Aditya Dhir

Applied AI Engineer · Production LLM systems & agentic workflows

#OpenToWork

About

Applied AI engineer with 5+ years building production LLM systems and multi-agent workflows at enterprise scale. Deep hands-on experience across the AI stack: RAG and semantic search on Amazon Bedrock and OpenSearch, LangGraph and AutoGen multi-agent orchestration, agentic tool-calling (MCP), structured outputs and function calling, plus the LLMOps layer — evaluation, observability, and safety guardrails — that keeps AI products reliable and governed in production.

What I'm looking for

I am looking for AI Tech lead or Lead software Engineer role.

Experience

InterVision / NWN

AI Tech Lead / Senior AI Platform Engineer

InterVision / NWN

Jan 2024 – Present

Lead AI platform architecture for enterprise SaaS — built the LLM inference backend powering live AI products across 10+ microservices and a team of 3–5 engineers. Architected and shipped agentic AI workflows and multi-agent LLM automation in Python and Go. Built an Amazon Bedrock gateway exposing internal Lambda functions as MCP tools, enabling Bedrock-based AI agents to take real actions across enterprise ITSM systems — creating support tickets and requesting knowledge-base access. Designed AppSync, Lambda, and microservices backend infrastructure powering enterprise AI chatbots, integrating Amazon Bedrock (Claude) for grounded, tool-augmented responses. Implemented LLMOps best practices in production: evaluation frameworks, model/output monitoring, observability, and safety guardrails. Cut inference and API latency ~25% through query optimization, caching, and service-boundary tuning. Deployed containerized AI services on Docker/Kubernetes to enterprise uptime SLAs; mentored engineers and drove design reviews.

Amazon BedrockPythonGo

Founding Engineer

MetaKeep (Passbird Research Inc.)

Jan 2022 – Jan 2024

Founding engineer — owned the backend platform end to end, shipping to production in under 3 months at hundreds of thousands of requests/day. Designed high-throughput microservices; led data modeling across PostgreSQL, MongoDB, and DynamoDB. Cut API response latency ~30%; partnered with founders on architecture and scalability trade-offs.

PostgreSQLMongoDBDynamoDB

Education

Liverpool John Moores University (via upGrad)

MSc / PGP · Data Science

Certifications

Microsoft Azure AI and Agentic Apps Developer

Microsoft

Jun 2026 – No expiry

Skills

FastAPIGoPythonobservabilitysafety guardrailsmodel monitoringLLM evaluationfunction callingstructured outputsprompt engineeringsemantic searchRAG pipelinesMCP toolsmulti-agent orchestrationAnthropic APIOpenAI APILangChainAutoGenLangGraphAmazon Bedrock

Languages

English (Full professional proficiency)