At Smart Ecosystems Inc, I audit Python solutions against production-grade standards, improving AI-generated code accuracy by 20%. I also engineer prompts to expose weaknesses in model reasoning and help mitigate technical hallucinations.
As an AI Software Engineering Fellow at Headstarter, I scaled semantic search to 8k+ stocks using Hugging Face embeddings and Pinecone. I improved query relevance and preserved context across company profiles.
At Headstarter, I also maintained AI explanation availability with API timeouts and local fallback, and reduced regression bugs by 30% by automating tests and linting with Pytest and GitHub Actions.
As a Software Engineering Intern at Internpulse, I achieved 96% test coverage for REST APIs, authentication, and database models. I also built projects including a churn-prediction inference system and an automated stock research platform.
