AI Software Engineer in Silicon Valley. I build systems that think autonomously

linkedin · email · github

Agentary - Autonomous multi-agent research platform. Deploy crews of specialized agents that research any domain, make voice calls via Twilio + Pipecat, and generate structured reports. Not a wrapper around an API - actual orchestration with task delegation, vector retrieval, and 10 external connectors. Python FastAPI PostgreSQL Redis Qdrant Celery Gemini Twilio

Basketball Intelligence - Full-stack NBA analytics platform over 5.7M shots and 45 years of box scores. Natural-language queries against the dataset, plus YOLO-based film analysis that breaks down real game footage. Started as a weekend question, turned into a real system. TypeScript Next.js SQLite Python YOLO

Disaster Damage Assessment - Professor-led research. VLM-powered pipeline that transforms paired pre/post aerial imagery into structured damage classifications using prompt engineering, batched inference, and caching. Includes a geospatial Q&A system that converts natural language into PostGIS queries and evaluates against FEMA labels. Python VLM SQL/PostGIS GeoJSON React

SoundScore - Cross-platform music logging and social discovery app. Native Android (Kotlin/Jetpack Compose) and iOS (SwiftUI) clients backed by a shared TypeScript/Fastify API. Ratings, reviews, social feeds, curated lists, weekly recaps, and Gemini-powered recommendations. Kotlin SwiftUI TypeScript Fastify PostgreSQL Redis

ReverseRecruiter - AI-powered reverse recruiting system. Automates company research, role discovery, and hiring-manager identification. Integrates voice-based cold calling via Pipecat + Gemini Live + Twilio to contact businesses directly and surface opportunities with source-backed findings. Python FastAPI Next.js Celery Qdrant Twilio

AI Software Engineer at a Stealth Foundational AI Company - Palo Alto, CA (Dec 2025 - Present) Building foundational AI product infrastructure end-to-end. Shipping under real user and investor scrutiny.

ML Intern at Haystream (May - Sep 2023) Built an end-to-end credit-risk pipeline: GBDT models lifting AUC +6 pts, containerized FastAPI microservice for real-time scoring, Airflow ETL, Power BI drift monitoring.

Full-Stack Engineering Lead at UT Dallas EPICS (Jan - May 2025) Led a 5-engineer team building a Next.js + Prisma grant-matching portal. 1,500+ opportunities cataloged, p95 search latency < 90ms, serverless ETL on AWS Lambda ingesting 300+ sources.

I care about the full loop: ingestion, processing, interface. Most of my projects go from raw data to a working frontend because the interesting engineering lives in the seams between layers, not inside any single one.

languages     python  typescript  c/c++  java  kotlin  swift  sql
ai/ml         multi-agent orchestration  VLMs  YOLO  gemini  tensorflow  embeddings  edge inference
backend       fastapi  spring boot  express  flask  celery  airflow  kafka
frontend      next.js  react  angular  swiftui  jetpack compose
data          postgresql  redis  qdrant  dynamodb  elasticsearch  mongodb  sqlite  postgis
infra         aws  docker  kubernetes  mqtt  git  ci/cd

• On-device ML - pushing inference to the edge where latency matters more than accuracy ceilings. Built a YOLOv8 + TFLite pipeline sustaining 80 FPS on a Jetson Nano.
• Autonomous agent evaluation - how do you actually measure if an agent is doing good work? Unsolved problem I think about constantly.
• Quantum computing - early-stage research interest, exploring the computational models.

_{shipping from dallas · graduating may 2026 · open to interesting problems}