Persona
Real-time voice-AI for clinical pre-intake. A speech-to-speech web app where a non-technical collaborator can shape conversational "personas" for therapist and psychiatrist intake.
Have you met Nick?
Full-stack engineer, 8+ years building AI products at venture-backed startups. I prototyped GPT-3 during the 2020 closed beta and have been building LLM and agentic tooling ever since — owning products end to end, from interface to inference.
About
Nice to meet you.
8+ years shipping
AI products at venture-backed startups
Computational Linguistics at Yale
Reasons about models from the inside
Building with LLMs since 2020
GPT-3 closed beta, now agentic tooling
Builds outside of work
Three personal products, taken from idea to shipped
I'm a product-minded full-stack engineer. I ship across the whole stack — mobile, web, backend, and the AI layer — and I'm just as comfortable talking to users and stakeholders as I am in the codebase.
My training in computational linguistics lets me reason about model behavior from the inside, not just through an API. I've raised a pre-seed round as a technical co-founder, led search products serving 100k+ businesses, and put ML models into tools that moved tens of thousands of home transactions.
Selected work
Things I've built
Thread
AI mental-health companion on iOS (TestFlight beta). Voice journaling, mood tracking, a support-system layer, and a PDF report generator built for therapists.
Campaign Companion
Full-stack TTRPG campaign manager — session planning, real-time die rolls, and AI-generated NPC dialogue, in one tool for game masters.
GPT-3, 2020 closed beta
Early access during OpenAI's closed beta at Podium — built automated support templates, chatbots, and review responses, presented directly to the CEO.
Experience
Where I've worked
May 2025 — Now
Senior Software Engineer · Outsmart
Stealth edtech startup ($38M raised); free college-level content via mobile, launching 2026.
- Pitched leadership and won buy-in to rewrite an internal AI-powered course-content generation platform used by 10 content contractors, then delivered the full-stack migration from React Native Web to Next.js + shadcn in two months.
- Eliminated significant tech debt in the rewrite — cutting dead code by 30%, replacing thousand-line legacy components, and re-architecting the codebase to be AI-extensible.
- Building the product's first in-app agentic LLM chat experience, chosen after prototyping a Vellum-style workflow generator, MCP-driven exercise creation, and RAG pipelines.
- Built core React Native features ahead of launch: a custom math keyboard, drills/exercises engine, and notifications.
Apr 2023 — May 2025
Career Break & Independent Work
- Contributor to a pending USPTO patent for an ML/NLP-powered grammar search engine (language-learning application).
- Provided product and technical consulting to an early-stage startup — user research, MVP scoping, and feasibility assessment.
- Pursued a dedicated creative-writing practice, completing fiction and narrative-nonfiction workshops at Sackett Street Writers' Workshop, GrubStreet, and Longhand.
Mar 2022 — Mar 2023
Software Engineer, Full-Stack · Opendoor
- Implemented ML models into valuation tools, enabling the sale of 40,000 homes and acquisition of 35,000 in 2022.
- Led end-to-end development of an image-sourcing feature, reducing costs by $600,000 annually.
- Authored product requirements and integrated user feedback for an internal real-estate communication tool.
Feb 2020 — Aug 2021
Tech Lead, Search Products · Podium
- Pioneered prototyping of OpenAI's GPT-3 during the closed beta — automated support templates, chatbots, and review responses presented directly to the CEO.
- Advised a dozen-plus teams on search-product implementation, including a Slack-like messaging search spanning messages, attachments, and channels.
- Led a design system and React library for consistent in-app reporting, grounded in user research and personas.
Oct 2018 — Aug 2019
Software Engineer · Podium
- Established a dashboard design system and React library, driving cohesion across six product teams.
- Maintained a back-end reporting pipeline supporting hundreds of millions of records.
- Rebuilt Elasticsearch indices for a search product, cutting index size 75% and improving query times.
Jun 2019 — Feb 2020
CTO & Co-Founder · Kilter
Ultra-searchable car-shopping platform.
- Raised $600,000 in a pre-seed round and drove the technical vision for a car-search MVP with sticker-level search — owning product and engineering end-to-end in a zero-to-one environment.
Education
Where I studied
Yale University — B.A., Linguistics (Computational Focus), 2018
HackYale lead instructor — wrote and taught a four-week NLP course in Python to college and MBA students.
Stack & strengths
What I work with
AI / LLM
LLM apps · agentic systems · RAG · MCP · prompt engineering · real-time voice (speech-to-speech)
Languages & frameworks
TypeScript / JavaScript · Python · Node.js · React · React Native · Next.js · SQL · Elasticsearch
AI tooling
Agentic coding workflows — Claude Code (since mid-2025) · Codex
Product & collaboration
Customer research · A/B testing · technical leadership · stakeholder communication
Let's build something.
I'm looking for engineering roles at AI-native companies. If that's you, I'd love to talk.
