Work
Where I've worked.
Research labs, startups, and big tech.
AI Research Intern
May 2026 - Aug 2026Samsung Research America
- Built an on-device signal router for always-on ambient AI on Bixby Edge AI, filtering device signals on-phone at 95% suppression and 90% intent-routing accuracy to reserve frontier models for events that need them.
- Built the developer SDK for Samsung's cross-device agent orchestrator, first deployed on XR glasses, for registering custom tools and skills, and unified several prototypes into one orchestrator.
Machine Learning Engineering Intern
Jun 2025 - May 2026Peraton Labs
- Developed RL agent for IoT malware detection using PyTorch and graph neural networks, reducing exploration latency by 35% and increasing detection coverage by 25% across 500K+ daily device events.
- Built large-scale data analytics ETL pipelines and heterogeneous graph architecture with autoencoders to model device communication patterns, accelerating policy convergence by 40%.
Computer Vision Researcher
Feb 2025 - Aug 2025Memories.ai
- Architected video memory framework for AR applications processing 10K+ streams using Python, Flask, and PostgreSQL, achieving 60% throughput improvement through frame sampling optimization.
- Published Python SDK on PyPI with 2K+ downloads, implementing async processing and REST API integration for video analysis workflows.
Undergraduate Robotics Researcher
Mar 2025 - Jun 2025IDEAS Lab
- Built real-time SLAM pipeline in C++ and Python with sensor fusion and Kalman filtering, improving 3D reconstruction accuracy by 25% and reducing latency by 30%.
- Implemented neural radiance fields for novel view synthesis, generating photorealistic scene reconstructions for autonomous navigation.
Data Engineering Intern
Aug 2024 - Dec 2024AgRPA
- Built end-to-end weed detection pipeline processing 200GB+ drone imagery with Python and cloud storage, implementing parallel ETL workflows and optimized data partitioning for 40% faster analysis.
- Engineered segmentation models achieving 92% accuracy on 50K+ images through agile sprints, reducing herbicide usage by 60% and costs by $150K annually.
ML Science & Engineering Apprenticeship
Jun 2023 - Aug 2023Naval Research Laboratory
- Developed ML models improving underwater transmission loss prediction by 20% over physics-based simulations.
- Built LangChain RAG system for classified document retrieval, reducing query time by 65%.