Ai Engineer
- Constructed LLM prompts using modified-RAG, improving end-to-end pipeline accuracy by over 90%- Researched opensource LLMs such as Llama 3, Qwen, mistral-nemo. Proposed a fine-tuned T5 model and presented comparison results to HQ- Implemented Python training code to fine-tune models through hyperparameter tuning; Plotted training graphs to identify.