Summary
Prasanth Pagolu is an ML Data Engineer and LLM specialist with 11 years of experience building production-ready models and MLOps pipelines across finance, pharma, and enterprise sectors. He’s delivered measurable efficiency gains—e.g., architecting continual learning for LLaMA-3.2-3B that achieved 97% parameter efficiency and slashed per-phase training costs from $800 to $100—by optimizing multi-GPU training and data pipelines for low-resource languages. At Sage he designs end-to-end SageMaker systems including feature stores, LLM monitoring, semantic search, and Bedrock agent integrations, bridging research-grade model work with robust cloud engineering. He combines deep hands-on skills (PyTorch, LoRA/QLoRA, vLLM, NCCL, CUDA) with reproducible MLOps (W&B artifacts, S3, SageMaker) to take prototypes to scalable deployments. Unusually, he pairs multilingual therapeutic dialogue work for African mental health initiatives with a strong background in Java enterprise systems, making him adept at both bleeding-edge model engineering and pragmatic production constraints. Based in East Hertfordshire, UK, he holds advanced data science training from Cambridge and an MSc in IT, bringing both academic rigor and long-term delivery experience.
11 years of coding experience
6 years of employment as a software developer
MSc, IT, MSc, IT at University of Liverpool
Advanced Data Science Accelerator, Data Science, Advanced Data Science Accelerator, Data Science at University of Cambridge
BCA, Computer Applications, BCA, Computer Applications at Osmania University