Pod Quant / Machine Learning Engineer

**FIG** Pod Quant / Machine Learning Engineer; $250K–$350K base + 20% pod profit share + founding equity; New York, NY (Hybrid)

**About Company** Foundational Insurance Group (FIG) is an AI-native, full-stack specialty insurance carrier built to operate like a quantitative hedge fund rather than a traditional insurer. FIG owns its own balance sheet and organizes around small, autonomous product pods — each pod finds an underserved or entirely new insurance market, builds a differentiated underwriting strategy on proprietary data and ML pricing models, and runs its own book and P&L once cleared through FIG's internal research and validation gates.

FIG has raised $107M in seed capital. First products are expected to go live in under 90 days, with the business targeting cash-flow positive in year one and the first five products funded off balance sheet.

Founded by Benjamin Markoff (CEO) — who built and exited Founder Shield (insurtech, nine-figure gross written premium, acquired by The Baldwin Group in 2021), the tech-enabled MGA Scale Underwriting, and Broker Buddha (exited January 2023) — and Judah Sosnick (CTO), a founding engineer at an enterprise AI/prediction-modeling startup who went on to build a systematic equities trading platform and served as CTO of an entertainment holding company deploying LLMs and agentic AI. The team will stay intentionally small (tens, not hundreds, of employees), so every hire is load-bearing.

**About the Role** This role sits directly inside a product pod and owns that pod's risk model end to end — from sourcing differentiated data through pricing, backtesting, deployment, and live portfolio monitoring. The intellectual shape is quantitative research: find signal, evaluate risk, construct models, allocate capital, manage performance. The difference is the market — specialty insurance rather than equities.

**Responsibilities** - Identify and acquire differentiated, often non-obvious datasets for a novel or underserved insurance niche. - Build ingestion and data-quality pipelines, writing your own scraper or tooling when nothing exists yet. - Engineer features and predictive signals from acquired data. - Design, validate, and backtest pricing and underwriting models against proposed strategies. - Deploy models into production and monitor model and portfolio performance over time. - Refine pricing as claims and market data develop. - Connect technical modeling decisions directly to the pod's live P&L.

**Required Skills** A. Deep experience building large, powerful prediction models on large and varied datasets, from a quant research or applied ML background. B. Track record as a quant researcher at a quant fund, a researcher at an AI lab, or a machine learning engineer with strong model-architecture depth (e.g., from Google or AWS). C. Demonstrated creativity in sourcing and validating data — what exists, what would be useful, how to get it, and whether it actually contributes signal. D. Genuine independence — willing to build your own tooling, such as a scraper, rather than waiting on the platform team.

**Bonus Skills** - Direct experience in specialty or niche insurance, prediction markets, or adjacent alternative-data-driven research.

**Logistical Info** - Location: New York City, hybrid — flexible for A+ talent. Tri-State-based candidates: 2–3 days/week in-office. Candidates based elsewhere: roughly one week per month in NYC. Open to candidates relocating to the NYC area. - Compensation: $250K–$350K base + 20% profit share on the pod's profit above an agreed hurdle rate, + founding equity. Client notes a successful pod could put all-in comp in the seven-figure range. - Number of openings: Ongoing pipeline — client confirmed this seat can become a recurring hire as new pods launch. - **Other:** No visa sponsorship — candidates must already have US work authorization. Founding equity on this seat.

**Ideal Background** - Primary profile: quant researcher at a quant fund, or researcher at an AI lab. - Alternative profile: machine learning engineer with less quant or research pedigree but deep model-architecture experience — strong MLEs from Google, AWS, and similar translate well. - No fixed seniority bar on paper, but in practice the client has passed on candidates as "too green for our first quant hire" at ~3 years, with added concerns about "the overall stability of his background" (confirmed rejections, Aug 2026). Read this as: no floor on years, but a real bar on demonstrated quant depth and career stability — a PhD-fresh candidate needs to show that depth directly, not lean on the credential alone. - Example candidates the client flagged as an instant "yes": linkedin.com/in/gm5315, linkedin.com/in/colediamond.

**Green Flags** - Independent, contrarian thinker comfortable exploring products or markets that have never existed, or that legacy carriers abandoned after a blowup. - Treats data as a native language; genuinely enjoys the "is this actually signal" question. - Self-sufficient — builds what's missing rather than waiting on support.

**Red Flags** - Needs a mature data or tooling org to be productive; unwilling to build own scrapers or pipelines. - Only comfortable with well-trodden, already-validated datasets and problems. - No US work authorization and would require visa sponsorship — FIG does not sponsor. - Short stints at brand-name funds paired with backtested-only experience (no live trading) — client has flagged this combination as a credibility concern; verify depth before submitting. - Struggles to communicate technical reasoning clearly and thoroughly in conversation — two rejections cited this directly ("not a cultural fit, unable to answer questions in a thorough manner").

**Interview Process** 1. Interview with Benjamin Markoff (CEO). 2. Interview with Judah Sosnick (CTO). 3. Take-home project, with a 48-hour turnaround.