James Nicholson's CV
- Phone: +1 650 863 2300
- Email: nicholsonjf@gmail.com
- Location: San Francisco
- Website: nicholsonjf.com
Summary
Full-stack engineer who ships. 6+ years building production systems at Harvard, plus a recent streak at the AI frontier: hackathon placements across Anthropic / Qualcomm / Meta / PyTorch events, and consulting engineer for LM Studio. M.S. in Software Engineering, Harvard.
Projects and Highlights
Atrium
- Feb 2026
- Anthropic "Build with Opus 4.6" Hackathon
- Demo video
- Selected from 13,000+ applicants. Built a multi-agent technical hiring platform orchestrating three Claude agents (Architect, Observer, Copilot) behind a real-time swim-lane timeline UI.
- Next.js + TypeScript + Anthropic API.
Logentic
- March 2026
- 3rd place, Llama Lounge Agentic Hackathon
- LinkedIn recap
- AI agents replacing the manual chaos of freight brokerage: real-time notifications, AI-driven carrier negotiations, and a lightweight web app that meets offline carriers where they are.
- 3rd of 500 applicants in a 16-hour build. Demoed mid-hackathon to a real ICP who confirmed the negotiation agent solved a problem he hadn't yet articulated — and that he'd be a customer.
- Built the agentic system on CrewAI, agent-to-agent payments with Skyfire, third-party tool connections via Composio, and Snowflake as the AI-optimized data layer.
Hindsight
- 2025
- Local-first memory for macOS (LM Studio / Element Labs consulting engagement)
- github.com/nicholsonjf/hindsight
- Paid consulting engagement with LM Studio (Element Labs) to build a reference local-first agent on their TypeScript SDK. Shipped as Hindsight: passive screenshot capture plus on-device vision inference, building a searchable personal timeline.
- Swift capture daemon, React/Vite frontend, Express/SQLite/TypeScript backend. Delivered a public blog post and internal friction log that fed the SDK roadmap.
Beaker
- October 2025
- 2nd place, PyTorch / Qualcomm / Meta ExecuTorch Edge AI Hackathon
- LinkedIn Recap
- Voice-driven on-device lab assistant running Llama 1B on Snapdragon 8 Elite via ExecuTorch. React Native, fully offline.
HECTRE
- May 2024
- Harvard M.S. thesis, with Pfizer
- nicholsonjf.com/hectre
- Distributed, multi-threaded Python pipeline using Meta's Llama API to extract structured data from clinical-trial PDFs.
- 20× throughput improvement versus prior manual process.
Experience
Harvard T.H. Chan School of Public Health, Senior Software Engineer
- Sept 2019 – present
- Technical lead on a multi-year flagship website redesign; owned core architecture, mentored two engineers to full productivity, and led UAT across 12+ stakeholders covering 200+ test cases and 100+ bugs.
- Designed and shipped the SAML2 authentication/authorization layer integrating enterprise identity providers (Okta) across the Chan domain.
- Built an automated faculty-profile generator (PHP / JS / MySQL) and integrations across Emma (SSO, DKIM, contact sync), Salesforce, Snowflake, and WordPress multisite — replacing manual workflows with self-service tools.
- Earned M.S. in Software Engineering from Harvard concurrently, with an LLM-focused thesis.
Stanford Graduate School of Business, IT Systems Analyst
- Sept 2015 – Sept 2019
- Built and iterated a service-request platform (Python, Google Apps Script) that automated scheduling, metadata tracking, and class-capture integration — saved hundreds of staff hours per year.
- Shipped Tableau dashboards on platform data to drive management decisions.
- Supported high-visibility events including MBA Admissions webinars (400+ registrants) and global distance learning programs.
Skills
- Languages: TypeScript, JavaScript, Python, PHP, Scala, C, SQL
- Frontend: React, Next.js (App Router), Tailwind, React Native, Vite
- Backend: Node.js, Express, FastAPI, Postgres, Supabase, Redis, MySQL, SQLite
- AI: Anthropic & OpenAI APIs, Claude Code, MCP, multi-agent systems, LM Studio, ExecuTorch, CrewAI, Composio, local & edge inference
- Infra: AWS, Docker, GitHub Actions, CI/CD, Vercel, Okta / SAML2
Education
Harvard Extension School, MS in Software Engineering
- Thesis: LLM-based structured extraction from clinical-trial documents (with Pfizer).
- Coursework in AI, AWS, Operating Systems (C), Robotics/Autonomous Vehicles, Scala for Big Data, Mathematical Statistics, Cryptography.