Amit Kanderi

About Amit Kanderi

As a Senior GenAI Engineer, I specialize in solving the \"Last Mile\" challenges of AI adoption: accuracy, latency, and cost. I don\'t just wrap APIs; I design robust Agentic RAG architectures. My work focuses on moving from linear execution chains to cyclic, stateful graphs (using LangGraph) that enable AI agents to perform multi-step reasoning, self-correction, and dynamic tool selection.Currently at Tredence, I am architecting the next generation of RAG-as-a-Service platforms. My focus is on:Orchestration: Designing multi-agent systems that handle complex query decomposition and planning.Governance: Implementing strict guardrails for PII redaction, topic adherence, and hallucination mitigation using advanced evaluation frameworks.Performance: Optimizing inference latency and token costs through caching strategies and hybrid retrieval pipelines (Vector + Keyword + Re-ranking).My expertise is grounded in 6.5+ years of production Machine Learning experience. At dunnhumby and NEC Corporation, I engineered large-scale MLOps pipelines and NLP models (BERT, LSTM) for the Retail and Media sectors. This deep background in data engineering and traditional ML ensures that my GenAI solutions are built on solid engineering principles, not just hype.Orchestration: LangChain, LangGraph, AutoGen, LlamaIndex, Semantic Kernel.Models: GPT-4,GPT-5, Gemini (Vertex AI), DeepSeek, Llama 3, Azure OpenAI.Infrastructure: FastAPI, Docker, Kubernetes, Azure AI Search, Redis.Evaluation: Ragas, Arize Phoenix, TruLens.Let\'s connect to discuss building the next generation of intelligent, autonomous enterprise systems.

Experience

1. Senior Generative Ai Engineer

   Tredence Inc.

   Aug 2024 — Present · Gurugram, IN

   Driving safe, scalable GenAI adoption through governed platforms and high-accuracy RAG systems- Developed a governed multi-model orchestration backend via FastAPI, enabling seamless integration of GPT, Gemini, and Deepseek with policy guardrails for safe, enterprise-compliant usage- Architected and deployed a \"RAG-as-a-Service\" platform using Azure AI Search and Vector DBs, integrating pre- and post-retrieval optimizations (HyDE, re-ranking) that improved retrieval accuracy by 40%- Implemented self-correcting agentic RAG workflows using LangGraph for multi-step reasoning and dynamic query planning, significantly enhancing the autonomy and reliability of AI task orchestration- Led cross-functional teams to deliver production ML systems, successfully reducing development cycles by 40% while maintaining enterprise-level security standards.

Education

• KRISHNA INSTITUTE OF ENGINEERING AND TECHNOLOGY, GHAZIABAD

  Master of Computer Applications, Computer Science and Machine Learning

  2016 — 2019

• Maharaja Agrasen College

  Bachelor of Science -BSc(Electronics Hons.), Electrical and Electronics Engineering

  2012 — 2015

• Vidya Bharati School

  XII, Computer Science

  2010 — 2012