Senior Software Engineer — scalable AI infrastructure, production ML pipelines, and LLM/RAG development (SOMA AI Technologies Inc.)
Fine-tuningEngineered production machine learning workflows that involved fine-tuning and improving transformer-based language models on large labeled datasets. Implemented RAG and embedding pipelines to support knowledge retrieval, semantic search, and long-context reasoning for decentralized inference systems. Built automated evaluation and benchmarking to measure model quality, robustness, and validator scoring consistency. • Fine-tuned BERT/RoBERTa and open-source LLMs using millions of labeled samples • Developed RAG pipelines (chunking, tokenization, embeddings, vector retrieval) • Implemented semantic intelligence with transformer embeddings and similarity search • Created evaluation/benchmark frameworks for quality, robustness, and performance