Amit Kumar

About Amit Kumar

I am Data Scientist with a strong background in designing and implementing large-scale distributed ML and Analytical systems. My proficiency lies in successfully delivering end-to-end production-ready ML/Data Science projects, while also working on emerging areas such as Generative Al and LLM. With a hands-on approach, I have honed my skills in various machine learning techniques including Deep Learning, LLM, Linear & Logistic Regression, Regularization, Decision Tree, Random Forest, SVM, XGBoost, Dimensionality Reduction, NLP, Time Series Analysis, Anomaly Detection & Forecasting, and Computer Vision solutions. In addition, I am well-versed in visualization tools such as Tableau, matplotlib, and seaborn. My technical proficiency encompasses Machine Learning, Image processing, computer vision,Python, PyTorch, Tensorflow,and MLOps Principles. Furthermore, I have hands-on experience working with AWS Cloud and possess a fundamental understanding of Azure cloud. Throughout my career, I have successfully led teams and contributed to strategic, high-value Machine Learning & Data Analytical projects. I firmly believe in building STAR teams and fostering a culture of mentorship. As a passionate learner, I am always eager to embrace new challenges and stay at the forefront of the ever-evolving field of data science.

Experience

1. Senior Ai Ml Consultant

   EY

   Sep 2024 — Present · Bengaluru, IN

   Developed an AI-powered microservice using RAG + LLMs that automated risk assessments, cutting manual review time by 87.5%.Built and deployed a FastAPI microservice automating security & operational risk assessments, reducing manual review time by 87.5%.Developed a RAG pipeline (FAISS + Claude 3.5 Sonnet) to analyze project descriptions, detect risks, and recommend mitigation strategies.Designed hybrid synchronous/asynchronous APIs integrated with Jira webhooks, enabling real-time responsiveness and background analysis.Deployed with Docker + Kubernetes + Harness CI/CD, resolving complex firewall and outbound connectivity challenges.Stack: Python, FastAPI, SentenceTransformers, FAISS, Jira API, Docker, K8s, Harness, Nginx.Built Agentic AI agents with LangChain to automate Jira workflows via natural language, enabling real-time, context-aware automationBuilt Agentic AI agents using LangChain + LangFlow to automate Jira workflows through natural language conversations. The system classifies intent, extracts structured data, and performs ticket operations dynamically using ReAct-style orchestration. Deployed on AWS + Kubernetes, these agents enabled scalable, context-aware automation and real-time decision support for engineering workflows.Designed modular agent pipeline (Intent Classifier, Context Extractor, Question Generator) powered by GPT-4o for multi-turn, context-aware dialogues.Integrated with Jira REST API to dynamically create, retrieve, and delete tickets using LLM-based orchestration (ReAct).Deployed on AWS (EC2, RDS)+ Kubernetes + GitLab CI/CD, enabling scalable and resilient AI automation.Stack: Python, AWS, LangChain, LangGraph, GPT, LLaMA, RAG, Agentic AI, Docker, Kubernetes, Streamlit.

Education

• Centre for Development of Advanced Computing

  Master of Technology - MTech

  2020 — 2022

• ARMY PUBLIC SCHOOL FATEHGARH,FARRUKHABAD,UP

  High School

  2009 — 2010

• Lovely Professional University

  Master of Business Administration (M.B.A.)

  2017 — 2018

• Lovely Professional University

  Engineer’s Degree

  2013 — 2017

• ARMY PUBLIC SCHOOL FATEHGARH,FARRUKHABAD ,UP

  SENIOR SECONDARY SCHOOL

  2011 — 2012