Looking for a job

uday varma

@udayvarma

Message

AI/ML Engineer specializing in agentic AI and GenAI production systems, delivering reliable models end-to-end.

India

Message

What I'm looking for

I’m looking for a role where I can build agentic GenAI and MLOps systems that deliver in production—measuring accuracy/latency, deploying reliably, and partnering cross-functionally to tie AI outcomes to real business goals.

I design and ship end-to-end AI/ML solutions—from predictive models and classical ML pipelines to GenAI applications and fully autonomous agentic systems. With 3.8 years of hands-on experience, I focus on building AI that’s intelligent, reliable, and production-ready.

On the Agentic & GenAI side, I architect multi-agent systems using LangChain, LangGraph, and LangSmith. I build RAG pipelines and knowledge systems with FAISS, PgVector, ChromaDB, and graph databases—then deploy them across AWS Bedrock, SageMaker, Google VertexAI, and Azure OpenAI. My emphasis is on agents that can reason, plan, use tools, and operate autonomously across complex workflows.

On the Predictive ML & MLOps side, I productionize supervised and unsupervised models end-to-end. I handle feature engineering, training, fine-tuning, deployment, CI/CD pipelines, drift monitoring, and observability—treating MLOps as the foundation for trustworthy AI at scale. I also use feedback loops to monitor and evolve accuracy and latency alongside the software lifecycle.

I published research at NeurIPS 2025, introducing COBOLCodeBench—the first rigorous benchmark for evaluating 15+ LLMs on COBOL code generation—and I contributed the dataset and framework to Hugging Face and GitHub. In my work at Grid Dynamics, I built scalable GenAI pipelines for support automation with RAG, cutting manual intervention in issue resolution by 95%. Earlier, as an MLOps Engineer, I optimized enterprise agentic solutions and improved GPU utilization by 80% while reducing hosting costs by 66–74%.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

Grid Dynamics

Aug 2025 - Present (1 year)

Designed and deployed scalable GenAI pipelines to automate L1 and L2 support workflows using RAG, enabling context-aware, real-time technical assistance that reduced manual intervention by 95% in issue resolution.

Built intelligent triage and root cause analysis systems that parse support tickets in real time, auto-route complex issues to the right engineering teams, and surface actionable diagno

Generative AI Root Cause Analysis Knowledge Management Log Analytics GenAI Pipeline Monitoring

MLOps Engineer ( Contract )

Milvian Group

Mar 2025 - Aug 2025 (5 months)

Designed and deployed enterprise-grade Agentic AI solutions using LangChain, LangGraph, AWS Bedrock, and SageMaker — automating complex data retrieval, workflow orchestration, and multimodal document intelligence at scale.

Built automated SQL generation and natural language data analysis platforms; engineered multimodal document QA systems with RAG pipelines for context-aware enterprise decision

Generative AI Machine Learning MLOps Agentic AI Lang Chain AI Agents LangGraph

MLOps Engineer

Milvian Group

Mar 2025 - Jul 2025 (4 months)

Designed and deployed enterprise agentic AI solutions using LangChain/LangGraph with AWS Bedrock and SageMaker for workflow orchestration and multimodal document intelligence. Implemented MCP server integrations, optimized SageMaker multi-model endpoints for GPU utilization and hosting-cost reductions, and supported MLOps best practices for monitoring, versioning, and production reliability.

Generative AI Machine Learning MLOps Agentic AI Lang Chain LangGraph AI Agents AWS Sagemaker Model Context Protocol (MCP)Document Intelligence SQL Generation Tool Use Model Monitoring Model Versioning

Specialist Programmer (GenAI)

Infosys

Oct 2022 - Mar 2025 (2 years 5 months)

Architected and shipped RAG-based enterprise applications using Azure OpenAI and LangChain for internal knowledge management and client-facing AI products. Fine-tuned LLMs for domain-specific tasks to improve accuracy and reduce hallucinations, built tool-calling conversational agents, and contributed to LLM evaluation through a COBOL code-generation benchmark research effort.

Azure OpenAI Lang Chain Fine Tuning LLM Evaluation Benchmarking Enterprise Search Semantic Search Vector Databases OpenAI API Azure AI Transformers BERT PyTorch TensorFlow Flask REST API Development TruLens Neural Networks

Specialist programmer ( Generative AI developer )

Infosys

Oct 2022 - Mar 2025 (2 years 5 months)

Architected and shipped RAG-based enterprise applications using Azure OpenAI and LangChain, serving internal knowledge management and client-facing AI products.

Fine-tuned large language models (GPT, open-source LLMs) for domain-specific tasks — improving task accuracy and reducing hallucination rates in production deployments.

Designed and evaluated a novel COBOL code generation benchmark for L