Hire Sai K. - Fine-Tuning Engineer

Architect and scale end-to-end ML pipelines using Python and AWS, automating clinical workflows that improved decision accuracy by 30% for high-stakes healthcare use cases. Engineer optimized SQL queries to transform multi-terabyte clinical datasets into model-ready features, reducing data latency by 40% and accelerating cross-functional delivery. Optimize PyTorch architectures for sequence-based tasks, implementing advanced regularization techniques that boosted production stability by 25%. Deploy high-dimensional embedding indexes using Vector Databases (Pinecone/FAISS), enhancing inference reliability and semantic retrieval consistency in real-time environments. Streamline MLOps workflows using MLflow for experiment tracking and versioning, ensuring 100% reproducibility and seamless model promotion across production stages.

Developed enterprise-grade GenAI solutions using LangChain and LLMs, increasing knowledge discovery efficiency by 40% for complex financial data extraction. Fine-tuned Transformer models (BERT, LLaMA) using Hugging Face and PEFT (LoRA/QLoRA), adapting models to financial domains with a 25% gain in precision. Engineered RAG (Retrieval-Augmented Generation) workflows on AWS Bedrock and SageMaker, significantly reducing hallucinations and improving response relevance for stakeholders. Leveraged Azure OpenAI Service to implement semantic search on unstructured datasets, improving information discovery speed and relevance through metric-driven evaluation. Partnered with DevOps teams to containerize ML services using Docker and Kubernetes, monitoring inference costs and system latency to ensure 99.9% availability.