Machine Learning Intern — AI agent with TTS and LLaMA fine-tuning (CGI Inc.)
Built and deployed an end-to-end AI agent using text-to-speech capabilities with TTS. Fine-tuned a LLaMA 3B model to improve latency characteristics for downstream AI interactions. Converted the trained model to ONNX and implemented FastAPI authentication for enterprise integration. • Fine-tuning LLaMA 3B using Unsloth • End-to-end AI agent development on Hugging Face • Model conversion for deployment (ONNX) • FastAPI-based authentication design