Resume

Experience

Head of AI Apr 2025 — Present

HackerPulse · San Francisco

Designed and built the AI service that powers the product — engineering leaders chat with an AI agent connected to live data from GitHub, Jira, Slack, Notion, etc. to get answers about their engineering org
Built a self-reinforcing knowledge graph across all synced integrations — agents discover implicit relationships, each new connection helps surface the next, and the user-facing agent queries it via tool calls

Founding AI Engineer Apr 2023 — Dec 2024

Yuma AI (YC W23) · San Francisco

Fine-tuned Mistral 7B to replace GPT-4 for ticket classification, cutting per-call inference costs by ~100x. Created an LLM-as-judge evaluation system to safely validate prompt changes and migrate between models
Designed an active learning loop for ticket routing: embed tickets → SVM classifies known types → anomaly detector catches out-of-distribution tickets → LLM labels the unknowns → retrain
Implemented the retrieval pipeline combining embedding search with BM25-style keyword matching

Machine Learning Engineer Apr 2019 — Mar 2023

Triplebyte · San Francisco

Joined as the first ML hire; promoted to Tech Lead of Data & Infrastructure
Migrated candidate search from a Postgres read replica (20s on a bad day) to Elasticsearch (60–300ms). Enabled hybrid search: BM25, metadata filters, and vector similarity in one index
Trained a logistic regression to predict which candidates would respond to recruiter outreach — three features, doubled response rate from 20% to 40%
Built a recommendation engine for a two-sided recruiting marketplace using neural collaborative filtering: content-based features handle cold start, collaborative filtering takes over as interaction data accumulates

Data Scientist Oct 2016 — Apr 2018

B23 · McLean, VA

Built an ETL pipeline aggregating millions of points-of-interest from disparate sources for predictive foot-traffic analysis, sold to hedge funds
Automated transfer of hundreds of terabytes of imagery between AWS and Snowball Edge devices for the US Navy, replacing a process that involved burning thousands of DVDs

Software Engineer Jun 2013 — Oct 2016

Agilex / Accenture Federal Services · Chantilly, VA

Developed a semantic search application for matching intelligence reporting with collection requirements via latent semantic indexing, optimizing average query speed by 10x

MS, Data Science Georgetown University · 2016 — 2018

BS, Computer Science & Mathematics Virginia Tech, summa cum laude · 2009 — 2013 Phi Beta Kappa

Python, Ruby, TypeScript, SQL

PyTorch, JAX, NumPyro, OpenAI Agents SDK, LangChain, MCP, LoRA, Hugging Face, scikit-learn, XGBoost, pandas

PostgreSQL, FastAPI, Elasticsearch, Redis, AWS