Rohan Pandey

Building and scaling cloud infrastructure and developer tools at DigitalOcean.

• Systems design, reliability, and performance work across production services
• Developer experience improvements: APIs, tooling, and automation

Adaptive red-teaming and a CI gate for AI agent security — an attacker that evolves its own payloads from the agent's responses until it leaks a secret or takes an unauthorized action, then fails the build.

• Attack strategies adapted from published red-teaming research: Best-of-N, PAIR, TAP, Crescendo, and Rainbow Teaming
• Against four support agents hardened with a keyword denylist, a fixed suite broke 0/4; the adaptive attacker broke 4/4 in 6.8 queries on average
• Open core: MIT-licensed engine on PyPI (gauntlet-agent) that runs offline with no API key, plus a hosted dashboard for teams

Gauntlet site

Project-based course teaching AI-powered research — from formulating a question to submitting a paper.

• Structured workflow covering literature review, analysis, and writing
• Built for researchers who want to ship real publications

lumina-research.com

Building smarter recovery for model rocketry — guided pre-flight workflows, sub-400ms backup deployment, and shareable flight records for TARC & competition teams.

• Physics-engine flight simulation and real-time telemetry platform
• Designed for 4,000+ competition teams in a $240M global market

qorbitsystems.com

Designing a reinforcement learning framework to synthesize efficient arithmetic circuits for polynomials, achieving ~70% success on degree-m polynomials.

• PPO + GNNs with curriculum learning and symbolic verification; scaling to higher-degree synthesis with Monte Carlo Tree Search to improve sample efficiency
• Preparing open-source benchmarks and targeting submission to ICLR/ICML/NeurIPS

Built and deployed an end-to-end ML pipeline for large-scale text sentiment analysis, enhancing prediction accuracy through strategic model selection.

• Fine-tuned and evaluated transformer models on domain-specific data using custom scoring metrics
• Owned data preprocessing, model validation, and inference automation for production-ready workflows

Engineered high-performance Python data pipelines leveraging vectorization and parallel I/O to reduce latency; built a secure self-service web app for automated workflow configuration and monitoring.

Developed mathematical models of CAR T-cell and tumor interactions using systems of ODEs trained on patient-derived B-ALL data, applying regression and loss optimization to evaluate treatment efficacy.

Built a reinforcement learning framework with DDPG agents in MATLAB and Simulink, optimizing aerodynamic performance on axial turbomachinery simulations under complex physical dynamics.

Engineered a real-time eye-tracking system integrating EEG headset signals with deep learning to decode gaze to text, contributing to brain-computer interface (BCI) research.

Lucy & Artemis; led team of 11 on lunar ice mapping rover; PDR, budget, programmatics.

ARES: GPS-guided recovery, drogue logic, telemetry in C++/Python.