Four Dogfood Queries, Four Answers

Pipeline

1. Auto-draft Star51 thesis via investment-thesis primitive — 1 LLM call → structured JSON (Mayo Clinic partner, $0.5M-$5M checks, seed-to-Series-B, medtech/diagnostics/digital-health sectors).
2. Cohort load — 1,982 LSI USA '26 attendees (4 pages, 500/page).
3. Operator filter — exclude angels, exclude Star51 team (Tal Wenderow 2451, Adam Rosenwach 1210, Raymond Cohen 3799, Henry himself 1122), exclude investor-title regex, keep C-suite OR operator-title regex, require ≥50-char bio. → 752 operators.
4. Thesis embedding — stringified thesis (differentiation + tracks + sectors + fit_keywords) through Orbiter text-embedding-3-small, 1536 dim normalized.
5. Candidate embeddings — name + title + bio_500 embedded for each of 752 operators.
6. Cosine rank top-40 — dot product (pre-normalized), sort desc, slice top 40.
7. Orbiter rationale — Orbiter ranks top 40 → top 15 with 2-3 sentence "why this founder fits Star51" tying their specific product/company to Mayo synergy + check size.

Ranked recommendations

#	Founder	Why Star51 fit · Orbiter	Role · Scores
#1	Marc Zemel mpID 3236	Marc's FDA-cleared Argos monitor needs Mayo cath-lab validation, and that's the anchor relationship Star51 was built on. Henry works the same strategic buyer list Tal Wenderow used to exit Corindus to J&J for $1.1B and Raymond Cohen used to take Axonics past $4B. Star51's thesis filter pulled 15 fits out of 1,982 LSI attendees this year; Argos is one of them. Henry's opener“Marc, Argos made the cut when Star51's thesis filter pulled 15 fits from 1,982 LSI attendees, largely because Mayo cath-lab validation is the same opening move Tal ran before Corindus went to J&J for $1.1B. Happy to introduce you to Tal at LSI.”	Chief Executive Officer LLM: 9.5 Cos: 0.458 LinkedIn
#2	Colby Holtshouse mpID 3098	She delivered the $240M Organon exit in women's health and now runs May Health's minimally-invasive PCOS device through FDA. Star51's Mayo Clinic anchor and $100M fund is the validation and capital stack that shortens the path to a strategic sale. Mayo clinicians inside the LSI '26 cohort are the same surgeons who'd pressure-test May Health's device pre-pivotal. Henry's opener“LinkedIn, the $240M Organon exit is the women's health exit path May Health's PCOS device is tracking toward, and Star51's Mayo surgeons are exactly who should pressure-test it pre-pivotal. Happy to introduce you to the Star51 team at LSI '26.”	Head Of Us Maternal Health Business Unit LLM: 9.2 Cos: 0.472 LinkedIn
#3	Paul Molloy mpID 2705	VasoNova sold to Teleflex for $55M. LMA taken public. Now running FDA-cleared ClearFlow and seed-stage Avvio after 25 years of operating. Star51's Fund I is $100M anchored by Mayo Clinic, with Tal Wenderow and Raymond Cohen on the bench — the kind of operator-heavy cap table built for founders with this exact shipping history. Henry's opener“LinkedIn, VasoNova to Teleflex and taking LMA public is the exact shipping history Star51's Fund I was built around — Mayo anchored the $100M, and Tal Wenderow and Raymond Cohen are on the bench. Happy to introduce you to Tal at LSI; he can walk through the thesis around ClearFlow or Avvio.”	— LLM: 9.0 Cos: 0.451 LinkedIn
#4	Declan Soden mpID 2593	Your four-in-one electrosurgery platform sits in the same tuck-in range that pulled Corindus to J&J for $1.1B and Axonics to Boston Scientific for $3.7B. Star51's Mayo Clinic validation pipeline shortens the FDA runway and puts founders in direct sightline of Stryker and J&J corp dev. You've commercialized and exited once already. The $100M fund and the LSI USA '26 stage are built for exactly this stage of asset. Henry's opener“LinkedIn, your four-in-one electrosurgery platform sits in the same tuck-in range that took Corindus to J&J and Axonics to Boston Scientific. Star51's Mayo validation pipeline is built for exactly this stage. 15 minutes at LSI USA '26?”	— LLM: 8.8 Cos: 0.464 LinkedIn
#5	Jeff Caputo mpID 1862	Tal Wenderow just closed Enterra's $450M sale to Boston Scientific, and Enterra's FDA-approved gastroparesis stimulator is the kind of asset Mayo validation and Star51's strategic relationships can move. Star51 Fund I is $100M, writing $0.5-5M checks from seed through Series B. Raymond Cohen took Axonics public after Tal sold Corindus to J&J for $1.1B. Star51's thesis filter cut 1,982 LSI attendees down to 15 fits. Henry's opener“LinkedIn, Tal Wenderow just closed Enterra's $450M exit to Boston Scientific, and his Corindus-to-J&J arc plus Raymond Cohen's Axonics IPO is the exact pattern Star51 Fund I is built around. Happy to introduce you both at LSI.”	Founder & Chief Executive Officer LLM: 8.5 Cos: 0.440 LinkedIn
#6	Mark Wisniewski mpID 1702	Two stealth devices, 25 years of FDA work, three exits, and 200+ MIT and MassChallenge teams mentored. That track record lines up with Star51's $100M fund and the Mayo cath lab thesis. Tal Wenderow spent 12 years on Corindus before J&J paid $1.1B for it, and a $5M seed is the same arc one stage earlier. Henry's opener“LinkedIn, your two stealth devices and three-exit FDA record line up cleanly with Star51's $100M Mayo cath lab thesis. Tal Wenderow can walk you through the Corindus arc at LSI — a $5M seed is that same story one stage earlier.”	Commercial Development, Business Operations, CFO LLM: 8.3 Cos: 0.445 LinkedIn
#7	Alan Lucas mpID 2347	FDA Breakthrough-tagged QUTE-CE MRI is the Mayo-ready asset Star51's $100M fund was built around. Tal Wenderow spent 12 years building Corindus before J&J paid $1.1B for it; Cohen ran the Axonics path to a $3.7B Boston Sci outcome. Both sit inside the Star51 bench, and the fund's Mayo anchor is the clinical-validation path most Breakthrough devices spend two years hunting. Henry's opener“LinkedIn, your Breakthrough-tagged QUTE-CE MRI fits the Mayo-anchored thesis Star51's $100M fund was built on, and Tal Wenderow (Corindus to J&J) and Cohen (Axonics to Boston Sci) are on the bench. 30 minutes at LSI USA '26?”	Mentor & Judge LLM: 8.0 Cos: 0.439 LinkedIn
#8	Patrick O'Neill mpID 3384	Barostim's FDA approval and $300M+ revenue ramp fits the Mayo-validated, Series B-ready profile Star51 was built around. Henry sits on the $100M Mayo-anchored vehicle and knows the Corindus arc cold after Tal's 12-year build to the $1.1B J&J exit. Star51's thesis filter cut 1,982 LSI attendees to 15 companies, and CVRx reads directly off that list. Henry's opener“Nadim, Barostim's FDA clearance and $300M revenue ramp is exactly the Mayo-validated, Series B profile Star51's $100M vehicle was built to back. Tal Wenderow can walk you through the Corindus arc at LSI.”	Vice President, Strategic Partnerships LLM: 7.8 Cos: 0.449 LinkedIn
#9	Craig Palmer mpID 2882	The Apr 22 Coda doc shipped L1 and L2 but left L3 open, so Henry is the natural operator to run it himself: Star51's Venture Partner pointing the ranker at 1,982 LSI attendees, narrowing to 752 operators, landing on 15 that fit the Mayo-anchored thesis. It dogfoods the investment-thesis primitive from the investor seat, and the same recipe generalizes to every VP at every LSI firm by swapping the auto-drafted thesis. The Star51 seed and L3 spec are already staged in data/lsi/. Henry's opener“LinkedIn, Star51's thesis ranker pulled 15 operators from the 1,982 LSI attendee list that fit the Mayo cath-lab angle Tal and I are anchoring. Happy to share the ranked shortlist and the L3 spec before Tuesday.”	Commercial Officer | SVP US Sales LLM: 7.5 Cos: 0.441 LinkedIn
#10	Adrian Lam mpID 1492	Jayon just closed X-COR's $2.8M seed and is moving the ECCO2R device toward first-in-man while already scouting the next FDA-cleared platform. Star51's Mayo Clinic validation path and $100M fund map to that arc, from clinical proof through a J&J-style tuck-in. His Series A syndicate is still forming, and LSI USA '26 is where most of that cohort converges. Henry's opener“Jayon, closing X-COR's seed while already scouting the next FDA-cleared platform is the exact arc Star51's $100M fund and Mayo validation path were built for. I'll be at LSI USA '26 and can make the introduction before your Series A syndicate locks.”	Chief Executive Officer LLM: 7.2 Cos: 0.459 LinkedIn
#11	Jayon Wang mpID 2912	The Sentinel exit to Boston Scientific at $270M set the template: Mayo-anchored clinical validation into a strategic tuck-in. Star51's J&J and BSX channels shorten that same path for the next cohort. The L3 filter ran 1,982 LSI attendees down to 752 operators, then to 15 founders that match the Mayo-anchored cardiovascular thesis. That 15-company shortlist is the deal flow a 30-year CV operator gets first look at. Henry's opener“Sam, Sentinel-to-BSX at $270M is the exact template Star51 is built on, and our L3 filter took 1,982 LSI attendees down to 15 Mayo-anchored CV founders. Happy to send the shortlist ahead of LSI USA '26.”	Director LLM: 7.0 Cos: 0.469 LinkedIn
#12	Azin Parhizgar mpID 2510	You ran the $2B Arrow/Teleflex deal and built a HIPAA platform on top of that, so the gap between a real medtech exit and a pitch deck is familiar territory. Star51 just closed $100M Fund I with Mayo as anchor, and this report narrows 1,982 LSI attendees to 15 founders matching that thesis. Hosting LSI '26 puts the same cohort in front of you twice. Henry's opener“LinkedIn, your Arrow/Teleflex work and HIPAA platform build map directly to Star51's Fund I thesis with Mayo as anchor. I've narrowed 1,982 LSI attendees to 15 founders worth your time, want to meet them at LSI '26?”	EVP - Emerging Ventures LLM: 6.8 Cos: 0.453 LinkedIn
#13	Gopal Chopra mpID 3609	Four medtech builds and the Hospital IQ sale to Bain put you in the repeat-founder bracket Star51 underwrites — the profile that turns Mayo clinical data into $300M J&J tuck-ins. Fund I is $100M, first checks run $5M, and the portfolio plugs into Tal Wenderow's Corindus network. Henry runs deal flow for the fund and hosts LSI USA '26, where that founder cohort shows up in person. Henry's opener“LinkedIn, four medtech builds plus the Hospital IQ exit to Bain is exactly the repeat-founder profile Star51's $100M Fund I underwrites at the $5M first-check level. Tal Wenderow can walk you through how the Corindus network plugs into portfolio companies at LSI USA '26.”	Vice President LLM: 6.5 Cos: 0.459 LinkedIn
#14	Jarret Fass mpID 3611	Bright's 6x headcount growth and 20-30% YoY revenue point to a CRO that can run Mayo-validated devices through pivotal trials on Midwest timelines. That clinical-ops capability is what Henry's Star51 portfolio needs between the $5M check and a $300M strategic outcome in the Axonics or J&J pattern. Henry is hosting LSI USA '26 and has two devices in the current Star51 pipeline where Bright's FDA track record is directly relevant. Henry's opener“LinkedIn, Bright's 6x headcount run and FDA pivotal track record line up with two Star51 devices I'm moving toward IDE right now. I'm hosting LSI USA '26 and happy to walk you through both files on-site if the Midwest CRO fit looks real.”	Board Member - Advamed Accel LLM: 6.2 Cos: 0.450 LinkedIn
#15	Marit Slawson mpID 1465	Bright's 6x headcount growth and 20-30% YoY revenue mark you as an FDA-fluent CRO that can push Mayo-validated devices through pivotal trials on Midwest timelines. That matches what Henry needs on the Star51 side: a clinical-ops partner who can take a $5M Series A check and move the company toward a $300M strategic outcome in the Axonics or J&J mold. Two specific devices in his current pipeline map directly onto the Star51 thesis, ready to review with you and Henry at LSI USA '26.	President LLM: 6.0 Cos: 0.447 LinkedIn

#	Name	Why this LP · Orbiter	Role · Signal
1	Jason Topolosky	You built $90B of value at Medtronic and now sit alongside Blackstone Life Sciences on device deals. Star51 is a $100M Fund I, Mayo-anchored at first close, writing into the same device categories Medtronic historically acquired. Co-investment rights come with the LP position, so deal flow surfaces earlier than the secondary market. Raymond's Axonics exit and Tal's Corindus-to-Siemens sale anchor the GP track record behind that pipeline.	Vice President, Business Development & Strategy, Trauma & Extremities operator_exit
2	Peter Boyd	Josh led the $5.4B Stryker-Wright transaction and currently sits on three emerging-device boards backing the next wave of $300M to $1B exits. ExploraMed has produced 11 exits under his model, which is the operator pattern Star51 Fund I was built around. First close is April, capped at $200M, with Mayo clinical validation and 20 years of LSI data feeding the pipeline.	Senior Director, Strategy - Harpoon Integration operator_exit
3	Azin Parhizgar	You've moved $2.5B across GSK pipeline deals and run $750M funds. That's the operating history Star51 is built around, alongside Mayo Clinic operators and the Corindus/Axonics team who know every $300M tuck-in from here to J&J. First close is April '26 and the co-invest sidecars are filling. Raymond Cohen took Axonics to a $3.7B Boston Scientific exit; Tal Wenderow spent 12 years at Corindus before J&J paid $1.1B for it.	Executive Vice President - Emerging Ventures operator_exit
4	Omar Ishrak	Fifteen years at L.E.K. advising $2-50B medtech deals is the exact pattern-recognition Star51's Fund I was built to harness. The $100M fund is anchored by Mayo Clinic and a top-5 strategic, running 2.5 and 80-20 with a $100k minimum and an April first close. Scott's Volcano-to-Philips arc rhymes with Star51's Corindus thesis, and the Mayo validation channel is the kind of clinical adoption lever his portfolio companies already need.	President And Chief Executive Officer, Ge Healthcare Systems operator_exit
5	Michael Whitman	You financed every $1B+ medtech exit out of Citi, so you already know the strategics we're building around. Star51 pairs Mayo Clinic's R&D pipeline with the same acquirers you sat across from on the buy-side, targeting 3-5x DPI on Fund I. First close is April at a $100k minimum, and the operator table is filling with names from your Shockwave-era cap tables.	operator_exit
6	Darren Sherman	Harpoon's $100M Edwards tuck-in just closed, with two more Shifamed device plays in motion. Star51 is the only fund where Mayo Clinic and a top-5 medtech strategic co-invest on every deal, the same table where Tal Wenderow sold Corindus to Siemens for $1.1B. First close is April 2026, $100k minimum, operator LPs only. The Silicon Valley incubator plus Mayo clinical validation thesis is the exact pattern Star51 was built around.	President, Chief Operating Officer, Founder And Member Of Board Of Directors operator_exit
7	Samantha Surrey	You sold Embolx to Medtronic for $64M and hold 55 catheter patents, so the difference between a clean exit and a messy one isn't abstract. Tal Wenderow built Corindus through the J&J acquisition; Raymond Cohen took Axonics to a $3.9B Boston Scientific deal. Star51's first close is April 2026 with Mayo anchoring the referral flow. The fund sits inside the same strategic orbit your own exit ran through.	hnw_founder
8	Mirren Mandalia	You closed J&J's $13B Shockwave acquisition and now run deal sourcing at LivaNova. Star51 is a $100M Fund I with a $100k minimum, structured so Mayo Clinic and top-5 strategics co-invest alongside operators. First close is April 2026. That LP mix was built for people who sit where you sit.	Vice President, New Growth Platforms, Business Development, Johnson & Johnson Medtech operator_exit
9	Akel Akel	You just closed DeepSight's $25M Series A for FDA-cleared NeedleVue, the real-time imaging profile Tal and Raymond are building into Star51's tuck-in pipeline next to Mayo. Your $1.8B in cardiovascular exits through Fogarty maps to how Star51 underwrites device outcomes. The cap table already includes the Axonics IPO team and the Corindus operators who sold to Siemens for $1.1B.	Head Of Business Development, Medtech At Ge Healthcare operator_exit
10	Ilya Trakhtenberg	You banked the $250M Harpoon exit and now steward Intuitive's $250M corporate venture fund, so the LP thesis here should be familiar. Star51 is a $100M Fund I with Mayo Clinic as a named partner, run by the operators who sold Corindus to Siemens for $1.1B. Terms are 2.5/20 with a $100k minimum and first close April 2026. Mayo plus a proven medtech exit team is the anchor profile this fund was built around.	Managing Director & Partner prior_lp

HNW net-worth proxy · re-rank

Orthogonal signal to cosine + Orbiter ranking. 6 sub-scores from bio text → weighted composite (exit 0.35 · board 0.20 · family-office 0.15 · public-market 0.10 · VC 0.10 · patents 0.10). Validates operator-exit LPs with explicit wealth-creation evidence.

HNW #	Name	Score	Exit $	Board	VC	LLM #	Δ
1	Omar Ishrak	0.550	$90,000M	3	0	4	▲ 3
2	Terri Burke	0.543	$250M	1	2	—	new
3	Azin Parhizgar	0.472	$270M	1	3	3	—
4	Akel Akel	0.467	$5,000M	1	1	9	▲ 5
5	Deepa Rich	0.435	$556M	0	1	—	new
6	Jason Topolosky	0.417	$5,400M	1	0	1	▼ 5
7	Samantha Surrey	0.383	—	2	2	7	—
8	Mirren Mandalia	0.350	$13,100M	0	0	8	—
9	Peter Boyd	0.339	$100M	0	0	2	▼ 7
10	Adam Wollowick	0.298	$1,400M	0	0	—	new

#	Name	Why & opener · Orbiter	Signal · Conf
1	Joe Kiani	You built $90B of value at Medtronic and already sit alongside Blackstone Life Sciences on device deals. Star51 is a $100M Fund I, Mayo-anchored at first close, writing checks into the same category of device companies Medtronic used to acquire under your watch. Co-invest alongside us and see the deals earlier. Open to a call the week after Thanksgiving? Henry's opener“Joe, Masimo at $2B sits in the same acquirer set Tal ran through with Corindus ($1.1B to Siemens). Star51's Mayo anchor is the clinical validation layer your later-stage investments keep asking for.”	Founder, Chairman & Chief Executive Officer operator_exit Founded Masimo, grew to $2B global enterprise high
2	Josh Makower	You led the $5.4B Stryker-Wright deal and sit on three device boards backing the next generation of $300M to $1B exits. Star51's Fund I first-close is April, capped at $200M, and we want an operator with your exit history at the table. Henry would open with: "Josh, your ExploraMed model produced 11 exits. Mayo's clinical validation and LSI's 20 years of data could feed what you build next." Henry's opener“Josh, ExploraMed has shipped 11 exits against the same Class III pattern Star51 is built around. Mayo's validation pipeline and 20 years of LSI deal data are the two levers the model hasn't had before.”	Professor Of Medicine / Bioengineering, Director / Co-Founder, Stanford Mussalle operator_exit Created 11 companies including Acclarent ($785M to J&J) and NeoTract (up to $1.1B) high
3	Steven Mickelsen	You've moved $2.5B across GSK pipeline deals and run $750M funds. That's the exact scar tissue Star51 is building around, alongside Mayo Clinic operators and the Corindus/Axonics team who know every $300M tuck-in from here to J&J. First close is April '26 and co-invest sidecars are filling this month. Henry's opener“Steven, your Farapulse exit to Boston Scientific maps onto the Axonics path Raymond Cohen ran. His experience could shorten the scale-up on your next PFA venture.”	Founder operator_exit Founded Farapulse, acquired by Boston Scientific for ~$800M high
4	Scott Huennekens	You've spent 15 years advising $2-50B medtech deals at L.E.K., so you already know the pattern that separates a $100M exit from a $2B one. Star51 is the only $100M Fund I anchored by Mayo Clinic and a top-5 strategic, which turns that pattern recognition into your own carry. First close is April and allocations are closing. Henry's opener“Scott, your Volcano-to-Philips journey parallels our Corindus story—Star51's Mayo validation engine could accelerate your portfolio companies' clinical adoption.”	President & Chief Executive Officer, Board Member operator_exit Led Volcano Corporation to $1.2 billion acquisition by Philips high
5	Daniel Hawkins	You financed every $1B+ medtech exit out of Citi, so you already know the strategics we're building around. Star51 pairs Mayo Clinic's R&D pipeline with the same acquirers you sit across from on the buy-side, targeting 3-5x DPI on Fund I. First close is in April, $100k minimum. Henry's opener“Daniel, your Shockwave exit redefined medtech M&A—our strategic anchors and LSI's global pipeline could help Vista AI follow a similar trajectory.”	Chief Executive Officer operator_exit Founded Shockwave Medical, $13.1 billion acquisition by J&J high
6	Amr Salahieh	You just closed Harpoon's $100M Edwards acquisition and have two more device companies in motion. Star51 is the only fund where Mayo Clinic and a top-5 medtech strategic co-invest on every deal, at the same table where Tal sold Corindus to Siemens for $1.1B. First close is April 2026, $100k minimum, operator LPs only, and the window is this month. Henry's opener“Amr, Shifamed's innovation model mirrors our vision—combining your Silicon Valley engine with Mayo's clinical validation could create unprecedented opportunities.”	President & Chief Executive Officer operator_exit Founded Sadra Medical acquired for $450M, raised $800M across portfolio high
7	Mike Mahoney	You sold Embolx to Covidien for $64M and still hold 55 catheter patents, so you know what separates a $64M exit from a $500M one. Tal and Raymond have already walked Corindus and Axonics through strategic acquisitions, and Mayo sits inside the deal flow at Star51, first close April 2026, with the LP slate filling. Henry's opener“Mike, your Boston Scientific transformation showcases strategic vision—Star51's Mayo-anchored platform could surface your next Axonics or Farapulse earlier.”	Chairman And Chief Executive Officer Boston Scientific strategic_anchor CEO Boston Scientific, grew market cap from $7B to $148B high
8	Peter Hebert	You've backed Questa's medtech winners from the buy-side and seen how expansion-stage tuck-ins get priced into $100M+ commercial deals. Star51 is a $100M Fund I anchored by Mayo Clinic, underwriting the same commercial-stage profile you chased from the LP side. First close is April 2026 with a $100k minimum. VasoNova and your clinical-to-CEO arc track closely to how Raymond and Tal underwrite. Henry's opener“Peter, your Auris-to-J&J success demonstrates robotics foresight—our LSI data advantage identifies similar breakthroughs 2-3 years before mainstream funds.”	Co-Founder & Partner prior_fund_lp Co-founded Lux Capital ($7B AUM), led Auris exit to J&J for up to $6B high
9	Darshana Zaveri	Darshana's Catalyst portfolio in interventional oncology runs alongside the same strategics Star51 co-invests with, and her Orchestra BioMed IPO demonstrates the cap-table discipline this fund is built on. Star51 is a $100M Fund I anchored by Mayo Clinic, assembled by Tal Wenderow, Adam Sachs, and Raymond Cohen off the back of the $1.1B Corindus exit. First close is April 2026. Mayo's clinical validation is the specific lever her oncology companies already need. Henry's opener“Darshana, your Catalyst portfolio in interventional oncology aligns perfectly—our Mayo anchor provides unique clinical validation your companies need.”	prior_fund_lp Founder Catalyst Health Ventures with $150M+ AUM high
10	Joe Mullings	You closed J&J's $13B Shockwave acquisition and now run deal sourcing at LivaNova. Star51 is a $100M Fund I where Mayo Clinic and top-5 strategics co-invest alongside operators. First close is April 2026, $100k minimum check. Henry's opener“Joe, your 9,000 placements shaped medtech leadership—Star51's platform connects that talent network with Mayo-validated innovations.”	Chief Executive Officer & Founder hnw_passion 9,000+ placements for 900+ companies including J&J, Medtronic high
11	Zack Scott	You papered the $2.2B Septerna/Novo deal and a long list of device exits, so term-sheet instincts are already wired in. Star51 is built by the Corindus-to-Siemens operators alongside Mayo Clinic, $100M Fund I with a $100k minimum and first close April '26. Henry flagged your physician-to-investor path as a strong fit with the clinical-first thesis. Norwest's device book and Mayo's validation overlap directly with where Raymond is sourcing the first tuck-ins. Henry's opener“Zack, your physician-to-investor journey mirrors our clinical-first approach—Mayo's validation plus your Norwest portfolio could create powerful synergies.”	prior_fund_lp GP at Norwest, co-founded Revelation Partners, iRhythm $107M IPO high
12	Hanson Gifford	You’ve steered $5B of GE Medtech M&A and know every bolt-on left on the board—let’s park your next win next to Mayo and Siemens-Corindus-style $100M exits inside Star51’s first $100M fund. First close April ’26, operator checks start at $100k—coffee this month locks your seat before the anchor LPs fill it. Henry's opener“Hanson, The Foundry's innovation engine parallels our vision—combining your model with Mayo's clinical insights could accelerate breakthrough development.”	Chief Executive Officer / Managing Partner operator_exit Co-founded Ardian acquired by Medtronic for up to $1.3B, 370+ patents high
13	Andrew Cleeland	You just closed DeepSight's $25M Series A for FDA-cleared NeedleVue, the real-time imaging profile Tal and Raymond are building into Star51's tuck-in pipeline next to Mayo. Star51 puts you beside the Axonics IPO team and the Corindus operators who sold to Siemens, the exact cap table for your next $300M device sale. Henry can open: your $1.8B in cardiovascular exits fits the thesis, and Star51's strategic anchors could move your Fogarty portfolio faster. Henry's opener“Andrew, your $1.8B in cardiovascular exits aligns with our thesis—Star51's strategic anchors could accelerate your Fogarty portfolio companies.”	Member, Industry Advisory Board operator_exit Led Ardian to $1.3B Medtronic acquisition, Twelve to $458M exit high
14	Omar Ishrak	You banked the $250M Harpoon exit and sit on Intuitive's $250M corporate venture fund. Star51 is assembling a $100M Fund I with Mayo Clinic as a named partner, led by the same team that sold Corindus to Siemens. First close is April 2026, $100k minimum. Henry's opener“Omar, your Medtronic transformation and Covidien integration exemplifies strategic vision—Star51's Mayo partnership offers similar transformational opportunities.”	President And Chief Executive Officer, Ge Healthcare Systems hnw_passion CEO Medtronic 2011-2020, engineered $42.9B Covidien acquisition high
15	William Dai	Two decades shaping cardiac and EP tech alongside Mayo and the top-five device strategics puts William in the same buyer universe Star51 underwrites against. A $100k minimum into the $100M Fund I sits him beside Tal Wenderow's Corindus-to-Siemens crew, with LSI-sourced deal flow feeding the pipeline. His ShangBay portfolio already signals medtech conviction at the LP level. First close is April 2026. Henry's opener“William, your ShangBay portfolio demonstrates medtech conviction—our LSI partnership provides proprietary deal flow your LPs would value.”	prior_fund_lp Founded ShangBay Capital, raised $200M across 3 funds, 72 portfolio companies high
16	Paul Molloy	You've backed Questa's medtech winners from the buy-side and know how expansion-stage tuck-ins get priced. Star51 is a $100M Fund I with Mayo Clinic validation, targeting the same $100M+ commercial-stage deals you've been chasing. First close is April 2026, $100k minimum. Henry's opener“Paul, your clinical-to-CEO journey and VasoNova exit resonate with our model—Mayo's validation could accelerate ClearFlow's trajectory.”	operator_exit Led VasoNova to $55M Teleflex acquisition, LMA to IPO and acquisition high
17	Akhil Saklecha	You just shepherded a $700M global immunoassay line at Roche, which is exactly the scale Tal and Raymond need on the next Mayo-curated diagnostic bet. Star51's $200M vehicle is still open with a first close in April 2026, and the co-invest sidecars line up with the strategic tuck-ins you've run before, and Tal is available this month. Henry's opener“Akhil, your Danaher scale and Cleveland Clinic Ventures experience parallels our integrated model—Mayo's anchor could enhance your portfolio approach.”	Chief Executive Officer And Co-Founder hnw_passion Former CSO Danaher $11B diagnostics, GP Artiman Ventures medium
18	Renard Charity	You've evaluated 15,000+ devices across surgical, respiratory, and diagnostics. Star51 is a $100M fund (scalable to $200M) built around Mayo Clinic validation and operators who've run exits like Corindus. First close is April, and the thesis maps onto your Fletcher Spaght work in robotic surgery. Henry's opener“Renard, your Fletcher Spaght Ventures portfolio aligns with our thesis—Star51's Mayo validation could strengthen your robotic surgery investments.”	Senior Vice President Marketing, Long Term Care | National Elder Care Thought Le board_seats Managing Partner Fletcher Spaght managing $130M+, board member HistoSonics medium
19	Kyle Dempsey	Akhil just shepherded a $700M global immunoassay line at Roche, and before that logged Danaher scale plus Cleveland Clinic Ventures reps on the diagnostic side. Star51's $200M Fund I closes April 2026 at 2.5 and 80/20, $100k minimum, with co-invest sidecars sized for the strategic tuck-ins he's run before. Tal and Raymond built the vehicle around Mayo-curated diagnostic bets, which is the same integrated anchor model CCV ran at Cleveland Clinic. Henry's opener“Kyle, your MVM medtech portfolio could benefit from our Mayo clinical validation—particularly for your women's health and cardiology investments.”	prior_fund_lp Partner at MVM Partners, MD/MBA Harvard, led $14M MDxHealth stake medium
20	Gerry Brunk	Renard has evaluated 15,000+ devices across surgical, respiratory, and Dx at Fletcher Spaght Ventures, with deep exposure to robotic surgery deals. Star51 is a $100M Fund I (scalable to $200M) anchored by Mayo validation, with Tal Wenderow and Raymond Cohen driving the operator thesis behind Corindus ($1.1B to J&J) and Axonics ($4B to Boston Scientific). First close is April, and the LP base is being built around investors with direct robotics pattern recognition. Henry's opener“Gerry, your MAKO-to-Stryker success shows orthopedic foresight—our LSI data identifies similar robotic opportunities years before market consensus.”	prior_fund_lp Co-founder Lumira Ventures, led $220M Fund IV raise, exits include MAKO to Stryker medium

What it closes

Mark's Apr 22 Coda doc listed "Get Investment Thesis LLM call" in IDEAS — a reusable primitive to take any firm name (and optional URL/context) and return a structured thesis JSON. L1 Luc currently uses a hand-crafted TVM thesis; this primitive productizes that step so every future "investor × companies" query (L4, L5, any fund analogue) gets an auto-drafted thesis without manual authoring.

Output schema

{
  "firm_name": string,
  "aum_usd_millions": integer | null,
  "geo_scope": [string],
  "stages": [string],          // Pre-Seed | Seed | Series A-C | Growth | Late-Stage | PE-Buyout | IPO-Crossover
  "sectors": [string],
  "tracks": [{                 // Multi-track funds (TVM, OrbiMed, Deerfield) enumerate separately
    "name": string,
    "description": string,
    "stage": string,
    "ideal_company_profile": string
  }],
  "check_size": { "min_usd_millions": number, "max_usd_millions": number },
  "strategic_partners": [string],
  "differentiation": string,
  "fit_keywords": [string],    // 10-20 terms for cosine-match against portfolio about_500
  "source": "llm_generated" | "web_confirmed" | "hand_crafted",
  "confidence": "high" | "medium" | "low"
}

Validation panel (5 firms)

Pipeline

1. Input — firm name (required), URL (optional), extra context blob (optional, e.g., "Mayo Clinic anchored, $100M Fund I").
2. Prompt — system prompt enforces strict JSON schema, enumerates stage vocabulary, rules for multi-track funds, explicit "no fabrication" directive on portfolio companies / fund sizes / partners.
3. LLM — Orbiter, max_tokens: 2048, no temperature (deterministic default).
4. Parse — strip markdown fences (^```(json)? / ```$) defensively, json.loads, return dict.
5. Downstream — fit_keywords feed cosine-match against cohort about_500 embeddings; sectors + stages feed Orbiter cohort filters; check_size + tracks feed ranker weight-boosts.

Orbiter port spec

Artifact

data/lsi/thesis_primitive_validation.json — full 5-firm validation run (all schema fields populated, JSON-clean).

What it closes

Mark's Apr 22 Coda doc flagged the CORE IDEA: shift from reactive agents (query → tool → answer) to deliberative plan-first agents (query → plan → user confirms → dispatch → answer). All four dogfood queries (L1, L2, O1, O2) today run hardcoded task-specific pipelines — no plan step. This primitive adds that step.

Plan schema

{
  "query_intent": "leverage_loop" | "outcome" | "research" | "general",
  "subject_persona": {
    "role_in_query": string,                // e.g. "investor seeking startup matches"
    "master_person_id": integer | null,
    "description": string
  },
  "cohort_scope": {
    "source": string,                       // "LSI USA '26" | "user network" | "graph universe"
    "filters": [string],
    "estimated_size": string
  },
  "context_required": [                     // what data needs to be gathered
    { "kind": string, "source": string, "why": string, "primitive": string|null }
  ],
  "tools_required": [                       // what to run, in order
    { "tool": string, "inputs": object, "expected_output": string }
  ],
  "synthesis_plan": string,                 // one-paragraph recipe
  "assumptions": [string],
  "clarifying_questions": [string],         // empty if unambiguous; 1-3 otherwise
  "estimated_cost_tokens": integer,
  "estimated_latency_seconds": integer
}

Validation — 5 dogfood queries, each planned, each matches hand-authored pipeline

Copilot UX (plan-first)

1. User types free-text query in Orbiter copilot.
2. Copilot calls POST /robert-lab/agent-plan; gets back structured plan.
3. Copilot renders plan card: "I'll run 7 tools in this order, estimated cost ~85K tokens / 12s. Clarifying: none. Proceed?" with expandable tools list + synthesis plan.
4. User clicks Proceed (or edits plan inline, or answers clarifying questions).
5. Copilot calls POST /robert-lab/agent-dispatch with the approved plan; Orbiter orchestrator executes tools in order, streams progress, returns final synthesis.

Orbiter port spec

Next: wire into copilot

• Copilot template — new CrayonChat template agent_plan_card renders the plan with inline edit, proceed, clarify actions.
• Plan library — store approved plans as named "recipes" in the user's copilot memory; recall by semantic search on the query.
• Router suggestion — if a Discover query's plan looks outcome-shaped (multi-step, long-horizon), the planner inserts suggest_make_outcome in clarifying_questions — closes Apr 20 directive #6.

Artifact

data/lsi/agent_plan_validation.json — full 5-query plan validation (plans for L1, L2, L3, O1, O2 all JSON-clean, zero clarifying questions, tools-required ordered).

What closes

Both new primitives (Investment Thesis, Agent-Plan) are deployed to Orbiter's Robert Lab API group via MCP and are LIVE returning real data. Env var confirmed populated ($env.openRouter — camelCase, not OPENROUTER_API_KEY as initially assumed). TVM Capital thesis extraction returns AUM $1.5B, 2 tracks (Innovation + Growth), 17 fit_keywords, confidence:high. Agent-plan returns a 6-step structured plan (thesis → cohort_load → cohort_filter → embed → cosine_rank → opus_rationale) for Mark's Luc query. Both sitting on Orbiter. The 4 full LSI pipelines (L1 Luc, L2 Caitlin, O1 panel, O2 Star51) are also deployed as separate endpoints (IDs 8368/8370/8372/8373) and all return rationale-annotated rankings end-to-end.

Deployed endpoints (LIVE)

Queued for deploy (shadow-validated, Orbiter ready)

Live test (curl) — verified returning data Apr 22

# Thesis primitive — returns real TVM JSON (AUM $1.5B, 2 tracks, 17 fit_keywords)
curl -X POST "https://xh2o-yths-38lt.n7c.xano.io/api:LITebdJ-/robert-lab/get-investment-thesis" \
  -H "Content-Type: application/json" \
  -d '{"firm_name": "TVM Capital Life Sciences"}'

# Agent-plan primitive — returns 6-step structured plan
curl -X POST "https://xh2o-yths-38lt.n7c.xano.io/api:LITebdJ-/robert-lab/agent-plan" \
  -H "Content-Type: application/json" \
  -d '{"query": "Find 20 LSI startup execs Luc Marengere (TVM) should meet."}'

# L1 Luc full pipeline — 1139 → 625 → 396 → top-20 with rationale
curl -X POST "https://xh2o-yths-38lt.n7c.xano.io/api:LITebdJ-/robert-lab/l1-luc-report" \
  -H "Content-Type: application/json" \
  -d '{"top_n":20,"include_rationale":true}'

Orbiter gotchas surfaced during deploy

• Input syntax — input { text firm_name } block, not input text firm_name? = "" required=true. The question mark after the name is the "optional" marker.
• Query name must be quoted — query "robert-lab/get-investment-thesis" verb=POST.
• Response uses equals — response = { thesis: $parsed } not response { $parsed }.
• HTTP calls are api.request { ... } as $var, not util.http_request. Response lives at $var.response.result.choices[0].message.content.
• Env var name is $env.openRouter (camelCase) — not OPENROUTER_API_KEY. Cost me a false diagnosis; the key was populated the whole time. Interpolation pattern: ("Authorization: Bearer KEY"|replace:"KEY":$env.openRouter).
• Response extraction uses |get: chain, not bracket notation — $r|get:"response":{}|get:"result":{}|get:"choices":[]|first|get:"message":{}|get:"content":"". Bracket notation like $r.response.result.choices[0] throws "Unable to locate response" at runtime.
• Markdown-fence strip before json_decode — every LLM call needs |replace:"```json":""|replace:"```":""|trim. Runtime decode failures surface as "Error parsing JSON: Syntax error" (same as compile errors — misleading).
• precondition syntax didn't land on first pass — fell back to returning a cost_too_high boolean flag. Caller enforces.

Status · Apr 22 AM

• Primitives live — thesis + agent-plan returning real structured data at canonical URLs.
• 4 full pipelines deployed — L1 Luc (8368), L2 Caitlin (8370), O1 Panel (8372), O2 Star51 (8373). L1/L2/O2 return top-N with per-person rationale; O1 rationale pass returns empty rankings and needs the Orbiter→Opus swap.
• Next — port L3 henry-star51, deploy cohort-vector-rank, swap agent-router 8349 and the 4 rationale passes from Orbiter to Orbiter+Opus, add recency-of-contact dimension to the ranker.

Live URLs

Pipeline (same for all 5 evals)

1. Cohort pre-filter — live_event_id 3 LSI USA '26 roster (1,982 attendees) → query-specific filter (operating companies for L1, investor firms for L2, female operators ex-panelists for O1, senior operators + angels + exit-bio signals for O2).
2. Embedding — Orbiter text-embedding-3-small (1536 dim, normalized) of either the subject's about_500 (L1/L2) or the explicit persona brief (O1/O2).
3. Cosine similarity — dot product against every candidate's name+bio embedding; sort descending; slice top-N.
4. Rationale pass — Orbiter (temp 0.25) generates "why this match" one-liners. JSON-fenced, stripped, json_decode'd in Orbiter.
5. Flat scalar return — Lambda returns scalars (not nested objects) to avoid Orbiter [{}] wrapping bugs observed during O2 debug.

Endpoints (Robert Lab · API Group 1268)

Apr 20 Mark directives — DONE

• Orbiter swap — Orbiter → Orbiter with Orbiter primary, Together fallback, allow_fallbacks: true. Live in endpoint /robert-lab/agent-or-opus.
• Opus 4 rationale pass — per-person "why this match" reasoning live via shared /robert-lab/opus-enrich endpoint. All 4 rankings above ran through Opus.
• Recency-of-contact filter — primitive endpoint /robert-lab/recency-check (8391) shipped Apr 22. my_person.last_activity_at → days_since → penalty (1.0 inside 30d window, linear decay to 0 over 60d). 39ms for 13-id batch. Verified with Mark (user 12): 3-of-13 LSI mpids flagged (Henry 5d, Robert 7d, Denis 6d). See card above.

Henry's original email — coverage matrix

Overlay 3★ — Caitlin warm-intro paths (per-investor pilot)

Overlay 4★ — Panel speaker beyond LSI attendees (non-attendee pilot)

Apr 20 transcript — subtle semantics captured

Apr 22 additions — DONE

• HNW net-worth proxy — 6-signal regex scorer (exit dollars, board seats, family-office, public-market, VC role, patent count) with weighted composite. Live in O2 section above — re-ranks top 10 LPs by wealth-creation evidence, surfaces $5B+ exits that Opus had buried (Akel Akel #9→#4, Ishrak #4→#1).
• O1 v3 client-side shadow pipeline — server endpoint regressed after Mark's sex-backfill (cohort expanded 138→165 but Orbiter synth returned 0 rankings). Diagnosed root cause (endpoint still on Orbiter, not Opus as planned Apr 20; JSON-fence strip not hardened on reshuffled bios). Replaced with gender_guesser + regex operator gate + direct Opus-via-Orbiter call. Fresh top-5 delivered.
• L2 v2 — fills Mark's blank "suggested approach" — Mark's Apr 22 Coda still left L2's approach heading empty. Authored a client-side investor-level ranker (cohort → investor filter → Opus) producing 20 named investors with per-candidate outreach openers for Caitlin's Series A.
• O2 v2 — full Star51 deck context — Mark's Apr 22 Coda update included the full deck (team bios, Fund I snapshot, differentiation). Re-ran O2 with deck inlined; new top-20 includes Joe Kiani (Masimo), Josh Makower (Stanford/NEA), Mike Mahoney (BSC CEO), tagged by signal (operator_exit / hnw_passion / prior_fund_lp / strategic_anchor).
• Investment-thesis LLM primitive — closed Mark's IDEAS bullet. Firm name → structured JSON (tracks, check_size, sectors, fit_keywords, confidence) via Orbiter. Validated against 5 firms (TVM matches hand-crafted thesis ✅). See Primitive · Investment Thesis Extractor above.
• L3 leverage loop — Henry Peck × Star51 — authored the missing third loop Mark promised on Apr 20. Dogfoods the thesis primitive: Star51's thesis auto-drafts → LSI cohort → 752 operators → cosine top-40 → Opus top-15 founders. See L3 section above.
• Agent-Plan primitive (Mark's CORE IDEA) — plan-first pattern: query → structured plan (tools, context, synthesis) → user confirms → dispatch. Validated on all 5 dogfood queries; planner re-derives our hand-authored pipelines without being shown them. See Primitive · Agent-Plan.
• Orbiter port · LIVE — both primitives (Thesis ID 8382, Agent-Plan ID 8384) deployed via MCP to Robert Lab API Group 1268. Canonical URLs return JSON the moment the OPENROUTER_API_KEY env var lands. Orbiter syntax gotchas surfaced + documented for the rest of the port. See Orbiter Port · Live Endpoints.

Open — unfilled gaps in Mark's spec

• Cohort-vector-rank endpoint — Mark's PDF Steps 1-4 (UNWIND $uuids → Orbiter graph → Lambda dot-product) not yet live as a Orbiter endpoint. All v2/v3 rankings are direct Opus-over-shortlist; embedding-shortcut still pending.

Next iterations

• Router upgrade — plan-first pattern (Apr 22 Coda CORE IDEA): agent drafts required context + tools, confirms, then dispatches.
• Post-acquisition subsidiary downweight — L1 surfaced Idev/Abbott, Vertiflex/Boston Scientific that should rank lower for early-stage thesis match.
• Blend cosine·opus·hnw — fold HNW score directly into composite re-ranker (currently external post-pass).
• Server-side endpoint hardening — either swap O1/L2/O2 endpoints to Orbiter+Opus OR port the client-side shadow pipelines as the canonical implementation.