Built a multi-agent AI pipeline in Python, deployed on AWS EC2, that scores photos with Gemini 1.5 Flash and Cloud Vision API and generates videos with ComfyUI (RunComfy) for 200+ venues at 85%+ quality acceptance
Automated photo review with venue-specific scoring, saving an est. 15+ hours of manual checks on about 2,000 images
Turned business reports into animated videos with LangChain agents that summarize and script them in Manim
Teaching Assistant · Northeastern University
Jan 2026 – Present
Boston, MA
Principles of AI and Foundations of AI (Sept 2026 – Present): grading assignments and quizzes on Gradescope for undergraduate and graduate AI courses; holding 1-on-1 office hours
Algorithms, Graduate TA (May – Aug 2026): led interview-style oral exams on dynamic programming, greedy and graph algorithms
Theory of Computation, Lead TA (Jan – Apr 2026): managed TAs on rubric design and grading; held weekly office hours and 1-on-1 sessions for 128 students; created problem walkthrough videos
Skills
AI/ML & LLMs
PyTorch, RAG, ReAct agents, LangChain, OpenAI/Anthropic/Gemini APIs, TensorFlow, Claude Code