Reinforcement Learning Researcher

• ›Conducted multi-agent RL experiments focusing on reward shaping, curriculum learning, and training stability.
• ›Built reproducible training pipelines using W&B, structured configs, and automated scorecards.
• ›Automated distributed training infrastructure, improving reliability and experiment throughput.
• ›Analyzed rollout data to detect reward hacking, mode collapse, and agent identity drift.

ML Engineering Intern

• ›Engineered multi-agent LLM workflows deployed on GCP Kubernetes clusters.
• ›Fine-tuned LLMs with supervised objectives, optimizing learning rate schedules and evaluation metrics.
• ›Optimized infrastructure efficiency by 3x, resulting in a 50% reduction in cloud costs.
• ›Built ETL pipelines and analytics dashboards for benchmarking model performance.
• ›Developed evaluation metrics for retrieval-augmented generation (RAG) systems.

Software Engineer

• ›Designed custom PCB boards + embedded systems for rocket avionics (Teensy 4.1, Arduino, C++).
• ›Integrated UART, SPI, and I2C communication buses with RF modules, achieving reliable real-time telemetry with >95% packet delivery rate during test flights.
• ›Delivered algorithm optimizations, improving real-time system performance by 15%.

Software Engineering Intern

• ›Developed serverless ML pipelines using AWS Lambda for scalable inference.
• ›Deployed TensorFlow prediction models for real-time data processing.
• ›Implemented CI/CD automation to ensure reliability of machine learning deployments.