Built an end-to-end AI-powered learning platform using Python, React, FastAPI, PostgreSQL, embeddings, and NetworkX that converts natural-language career goals into personalized, prerequisite-ordered learning paths. Designed backend services for learner profiling, skill-gap analysis, recommendation, path generation, explanations, progress tracking, and feedback-driven adaptation. Implemented a directed acyclic graph and topological sorting to enforce prerequisite relationships and generate valid learning sequences. Built REST APIs and a React dashboard for onboarding, path exploration, progress tracking, and learner feedback, with automated tests for recommendation, skill-gap, and path-generation workflows.

Built a full-stack AI platform for RF signal threat classification using React.js, Django, FastAPI, PostgreSQL, PyTorch, ONNX, and TensorRT. Designed modular frontend, application, database, and ML inference services with clear service boundaries and REST APIs for authentication, user management, prediction workflows, and prediction history. Developed and tested end-to-end inference workflows, request validation, authentication, automated testing, and deployment components. Optimized trained models using ONNX and TensorRT for efficient inference and production-oriented deployment.