1. PHL2026 Comcast

   AI and DevOps Intern

   Built a team-wide AI agent portal and a 24/7 DevOps agent that reads logs, triages incidents, and ships gated fixes.

2. Remote2026 Handshake AI

   AI Fellow

   Prompt-evaluation pipelines and production fixes for code-assistant accuracy across releases.

3. Remote2025 Esri

   Software Engineering Intern

   Built a multi-agent onboarding assistant for ArcGIS migration, cutting onboarding time by 30% and errors by 40%.

4. PHL2025 Villanova CS

   AI Researcher

   Agentic research that simulates an investment bank. Analyst agents debate market signals, then a VP agent weights the final call.

5. PHL2024 Ernst & Young

   Technical Solutions Intern

   SAP S/4HANA migration support. SQL checks and Python scripts that cut data errors 15% and sped up batch processing 20%.

1. Asclepius

   HackPrinceton winner GitHub

   A privacy proxy between clinical trial staff and cloud AI models. Strips PHI, protects drug IP, and keeps the workflow fast.

2. Context Compass

   21 GitHub stars GitHub npm

   A context engine for Claude Code. It learns which code belongs together from git history and serves compact context over MCP; on a held-out Flask eval it lifts recall@10 by 31.9 points while using 35% fewer tokens.

3. AgentSAT

   GitHub Live

   Tests AI agents before users do. It builds company-specific test plans, stress-tests the real agent, turns failures into reproducible engineering reports, then retests the fixed version.

4. ColabCode

   Open source npm

   An open-source coding agent that runs open-weight models on your own Google Colab GPU. Your code stays on your machine; inference runs over an authenticated SSH tunnel.

5. Regime-Conditional Conformal Prediction

   Research GitHub

   Uncertainty for next-day stock volatility: prediction bands that keep their promised coverage within each market regime, not just on average, tested on ~394k out-of-sample stock-days.

Tier

Expert

Competitions

Rank

2,379

of 217,542

Silver

01

Medal

Bronze

02

Medals

1. NVIDIA Nemotron Model Reasoning Challenge

   Silver 169 / 4,185

   Reasoning techniques on NVIDIA’s open Nemotron models, scored on a new benchmark. Finished 169th of 4,185 teams.

2. The 2026 NeuroGolf Championship

   Bronze 248 / 2,963

   The smallest neural networks that still solve ARC-AGI image transformations. Finished 248th of 2,963 teams.

3. Orbit Wars

   Bronze 469 / 4,729

   A real-time strategy simulation: an agent that conquers planets orbiting a sun in 2- and 4-player matches. Finished 469th of 4,729 teams.

Also competed

• Pokémon TCG AI Battle
• ROGII Wellbore Geology Prediction
• March Machine Learning Mania 2026
• AI Agent Security

Competing now

• ARC Prize 2026 · ARC-AGI-2
• RSNA Knee Abnormality Detection
• Kaggriculture

Kaggle profile

Languages

• Python
• JavaScript
• TypeScript
• Java
• C++
• HTML5
• CSS3
• SQL

Frameworks

• React
• Node.js
• Express.js
• Tailwind CSS
• Bootstrap
• Django
• Flask

AI / ML

• LangChain
• LlamaIndex
• CrewAI
• Scikit-learn
• PyTorch
• NLTK
• TensorFlow

Tools & Cloud

• Git
• GitHub
• VS Code
• Docker
• AWS
• Google Cloud
• Azure
• Vercel
• Figma
• Postman
• Jira

Databases

• MongoDB
• PostgreSQL
• MySQL
• Redis