Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

AI/ML engineer with 3+ years of experience building and deploying production LLM systems, agentic workflows, and RAG pipelines for enterprise clients. Deep practitioner focused on transformer internals, retrieval quality, and reliable multi-step tool-using agents. Currently studying transformer internals at CMU Xu Lab and continuing peer-reviewed mechanistic interpretability research on why LLMs fail to surface correctly computed information. Previously built multilingual AI products from scratch and delivered production NLP/ML systems across classification, semantic search, and responsible model evaluation.…AI/ML engineer with 3+ years of experience building and deploying production LLM systems, agentic workflows, and RAG pipelines for enterprise clients. Deep practitioner focused on transformer internals, retrieval quality, and reliable multi-step tool-using agents. Currently studying transformer internals at CMU Xu Lab and continuing peer-reviewed mechanistic interpretability research on why LLMs fail to surface correctly computed information. Previously built multilingual AI products from scratch and delivered production NLP/ML systems across classification, semantic search, and responsible model evaluation.

Satwik Sunnam

AI Engineer, Programmer, Data Scientist, +2





AI/ML engineer with 3+ years of experience building and deploying production LLM systems, agentic workflows, and RAG pipelines for enterprise clients. Deep practitioner focused on transformer internals, retrieval quality, and reliable multi-step tool-using agents. Currently studying transformer internals at CMU Xu Lab and continuing peer-reviewed mechanistic interpretability research on why LLMs fail to surface correctly computed information. Previously built multilingual AI products from scratch and delivered production NLP/ML systems across classification, semantic search, and responsible model evaluation.…AI/ML engineer with 3+ years of experience building and deploying production LLM systems, agentic workflows, and RAG pipelines for enterprise clients. Deep practitioner focused on transformer internals, retrieval quality, and reliable multi-step tool-using agents. Currently studying transformer internals at CMU Xu Lab and continuing peer-reviewed mechanistic interpretability research on why LLMs fail to surface correctly computed information. Previously built multilingual AI products from scratch and delivered production NLP/ML systems across classification, semantic search, and responsible model evaluation.

Available to hire

AI/ML engineer with 3+ years of experience building and deploying production LLM systems, agentic workflows, and RAG pipelines for enterprise clients. Deep practitioner focused on transformer internals, retrieval quality, and reliable multi-step tool-using agents.

Currently studying transformer internals at CMU Xu Lab and continuing peer-reviewed mechanistic interpretability research on why LLMs fail to surface correctly computed information. Previously built multilingual AI products from scratch and delivered production NLP/ML systems across classification, semantic search, and responsible model evaluation.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Language

English

Advanced

Work Experience

AI/ML Engineer at East Rock Software

January 1, 2026 - Present

Developed an enterprise RAG knowledge assistant using LangChain, Hugging Face, and Pinecone for contextual retrieval and question answering. Built NLP pipelines for document classification, entity extraction, summarization, and semantic matching across internal business and customer-support content. Implemented LangGraph-based agent workflows with tool calling, validation steps, and fallback handling for reliable multi-step query processing. Created FastAPI microservices to expose LLM inference, embedding generation, retrieval, and response evaluation APIs. Containerized services with Docker and deployed scalable workloads on AWS with CI/CD for version-controlled releases. Established evaluation and monitoring across response relevance, hallucination risk, latency, prompt quality, and retrieval performance. Optimized LLM serving via quantization and latency improvements to reduce per-token inference cost at scale.

AI/ML Engineer at Fractal Analytics

January 1, 2021 - July 1, 2023

Designed ML and NLP solutions for text classification, sentiment analysis, intent detection, and customer-feedback analytics. Fine-tuned transformer models (BERT, RoBERTa, domain-specific variants) using PyTorch and Hugging Face with LoRA and full fine-tuning approaches to improve understanding of domain terminology. Enhanced semantic search using embeddings and FAISS for faster discovery in large unstructured document collections. Prepared and transformed data using Python, SQL, Pandas, and PySpark with reusable feature-engineering workflows. Built model training and validation pipelines using Scikit-learn, TensorFlow, and MLflow for experiment tracking and reproducible deployment. Integrated prediction services via Flask and REST APIs for secure business consumption. Applied explainability, bias assessment, error analysis, and performance monitoring to improve model reliability by 18% while supporting Responsible AI practices.