Hello world
Ankit Mukherjee
0
AWS Bedrock
Anthropic
LangGraph
FastAPIKubernetesWeaviateAWS BedrockAnthropicLangGraphFastAPIKubernetesWeaviate
Who I am
The engineer
behind
Ankit Mukherjee
I bridge robust enterprise architecture and high-velocity innovation, with a focus on distributed systems, AWS cloud infrastructure and applied AI. Today I'm an AI engineer at Pure Storage building production-grade agentic systems — LLM services, guardrails and evaluation, observable and secure on AWS. Before that: feature lead at AskTuring.ai, three years at PwC, and an MS in AI & ML from the University at Buffalo.
Not to be missed, worth sharing. Here's a pick of what I build.
Recent work
What I do
The toolbox
LLMs (OpenAI, Gemini, Anthropic), LangGraph, agentic RAG, RAGAS, PyTorch, CNNs, NLP and computer vision.
Python (FastAPI), Node.js (NestJS), TypeScript, Go, Java, C++ and SQL — built for real traffic.
Weaviate, pgvector, PostgreSQL, Redis, Kafka, Dramatiq and DynamoDB.
AWS (EKS, Lambda, Bedrock, CloudFormation), Docker, Kubernetes, GitHub Actions and CI/CD.
React, Redux, Next.js, JavaScript, HTML5/CSS3 and iOS development with Swift.
Prometheus, Grafana, Langfuse, PostHog, OpenTelemetry and CloudWatch.
Where I've worked
Experience
Building production-grade agentic AI systems that drive enterprise automation and digital transformation. Designing and deploying LLM-based services with prompt engineering, guardrails and model evaluation frameworks, delivered as secure, scalable, observable systems on AWS that plug into enterprise platforms.
Owned critical product initiatives from architecture and data modeling to production code. Led the Slack integration team, shipping an omni-channel enterprise chat experience from scratch. 12 API endpoints, 5 async workers and 100K+ messages ingested into Weaviate. A 44-node LangGraph pipeline with hybrid vector + BM25 search cut time-to-first-token by 40%, and an LLM-as-judge RAGAS framework lifted answer quality by 30%.
Spearheaded a GenAI initiative that secured a $12M+ client contract with an AWS Bedrock RAG pipeline. Tech lead for an enterprise platform used by 1,000+ people, built Kafka + DynamoDB microservices that cut latency 40% through 300% traffic spikes, and a Stripe revenue engine processing $1M+ a year.
Education
University at Buffalo (SUNY)
MS in Computer Science — Artificial Intelligence & Machine Learning · Aug 2024 – Dec 2025
Machine Learning · Deep Learning · Computer Vision · Distributed Systems · Data Intensive Computing
What colleagues say
Kind words
John“A pivotal member of the backend team — real-time chat over WebSockets, a scalable database design and AWS deployment.”
John P. Weiksnar — Tesla Owners Club NY State
Neslihan“Open, clear, and willing to bridge design and technology. The kind of teammate who elevates everyone's work.”
Neslihan Kilic — Conversational AI at SAP
Bhagya“Enthusiastic, adaptable and a quick learner with a natural knack for new tech. A true asset to any team.”
Bhagya Pasupureddy — Gen AI Manager @ PwC