Sean Carroll

Afoot · Cambridge, MA, USA

Building Afoot — a personal, geography-first list of things happening near you that you'd actually want to do. Discovery is an H3 hex, not a metro name: a demand-gated Gemini grounded-search sweep fills verified inventory, a preference vector learns from thumbs, and opening the app composes a curated track with light push nudges. Designing the extraction stack around SemDAG — a verified operator graph compiled from clustered model traces — so repeated event layouts drop off the large model onto cheap Python nodes, with a small leftover-field call only when the graph cannot cover a field.

• ·Pull-first tracks: occupancy-gated, budget-capped hex sweeps; every event must carry a live source URL and pass a page-fetch verification gate (schema.org Event as evidence, the model as a lead) before it can be served
• ·Preference scoring and slot composition on FastAPI + Postgres (H3 as B-tree text, no PostGIS); Expo app for onboarding, tracks, and going/ratings; Next.js public share pages
• ·SemDAG as the compiler rung for event extraction: weave verified operator DAGs from successful grounded-search traces so the standing sweep gets cheaper and more deterministic as layouts repeat — designed now, implementation next

SwarmHire · Cambridge, MA, USA

Building SwarmHire — a multi-agent career platform that scores a resume against a role the way a hiring team would, then stays useful through cover letters, job discovery, and application readiness. Six Hive specialists review in parallel; a Navigator ranks what to fix first. Owned the full stack from FastAPI/HiveAI agents through a Next.js product surface with Stripe billing.

• ·Hive Score: six-agent parallel review (ATS, achievements, completeness, market alignment, narrative) with inline accept/dismiss suggestions — not a single-model résumé grade
• ·Cover Letter Studio (generate, Tiptap refine, export), Hirebase-backed job library/CRM, salary intelligence, and application-readiness briefs so the score turns into a weekly search workflow
• ·Production loop: FastAPI + Supabase/Postgres, Next.js 16 + Tailwind, HiveAI/LangGraph agents, Redis, Stripe subscriptions and entitlements — built to ship, not demo

Stealth Startup · Cambridge, MA, USA

Building the AI transaction layer for the U.S. boat market in stealth — aggregating fragmented inventory into canonical listings, explainable deal scoring, and agent-assisted buyer workflows that connect qualified buyers with brokers.

• ·Multi-source inventory aggregation with comps-backed pricing on every listing
• ·Conversational buying agent that qualifies buyers and routes exclusive introductions to matched brokers

Freelance · Remote

Fractional CTO and hands-on technical partner for AI-forward startups from idea stage through Series B-scale pressure. Embed with founders and product teams to turn ambiguous requirements into shippable systems: technical discovery, architecture, build-or-buy decisions, vendor and model selection, and the people/process work (hiring rubrics, code review culture, CI/CD) that keeps velocity honest after the first demo.

• ·Portfolio spans $0–$25M ARR across home health care, real estate, e-commerce, and social — same playbook: map the constraint (compliance, data quality, latency, cost), design the smallest production slice, and ship with observability from day one
• ·Deliver LangGraph (and adjacent) agent workflows, RAG over messy domain corpora, eval harnesses, and LLM-backed APIs that respect tenancy, PII, and each client's infra maturity — usually Python/FastAPI services paired with Next.js or existing app surfaces
• ·Act as the single thread between execs, domain experts, and engineers: roadmap facilitation, sprint-level prioritization, interview loops, and crisp written specs so teams do not stall on alignment

Partnering with startups and growth-stage teams for fractional CTO and strategic technical leadership. Open to selective new engagements.

Learn more about CTOaaS