Hello, I'm

|

|

Specializing in Machine Learning, NLP, Generative AI & Agentic AI. Proven ability to build scalable pipelines, derive actionable insights, and architect intelligent solutions from end-to-end.

Sachin Mishra
Gen AI
Analytics

Key Achievements & Open Source

locdataMAC (PyPI)

Open Source Creator

Authored and launched locdataMAC, an open-source Python package on PyPI. It acts as a comprehensive productivity pipeline that tracks user activity across various software applications and websites, providing rich analytics on time utilization and digital habits.

google-okf (PyPI)

Open Source Creator

Created and published google-okf, an open-source Python library designed to generate Google Open Knowledge Format (OKF) bundles from various data sources. OKF standardizes knowledge sharing structures to optimize data exchange protocols.

Gold Medalist

Awarded the prestigious Gold Medal out of numerous participants for outstanding performance, innovative thinking, and technical excellence during the "IC the future" academic competition.

About Me

I am a results-oriented Data Scientist with 5+ years of experience in developing and deploying cutting-edge machine learning solutions. My passion lies in solving complex problems using GenAI, Agentic AI, Model Context Protocol (MCP), A2A Protocol, Machine Learning, Natural Language Processing (NLP), and Computer Vision.

I thrive on building scalable models and productionizing AI applications using cloud platforms like AWS, Azure, and GCP. From conceptualizing Local RAG systems to orchestrating multi-AI agent workflows, I am dedicated to driving innovation and delivering tangible business impact.

5+

Years Experience

10+

AI Projects

3

Cloud Platforms

Education

Bachelor of Engineering

Electronics & Telecommunication

Thakur College of Engineering & Technology

CGPA: 9.04/10 - Gold Medalist for "IC the future" Competition

Professional Journey

Mar 2025 - Present

Data Scientist

Tata Technologies

Driving advanced AI initiatives and building end-to-end data pipelines for scalable machine learning deployments.

Aug 2024 - Jan 2025

Data Scientist

Employtech Center India Private Limited

Implemented a Local RAG System using a Mistral-7B quantized model, reducing assessment creation time by 2-3 man-days per course and improving productivity by 15%. Also developed the "assessment-creator" Python package published on PyPI.

Jul 2022 - Jul 2024

Data Scientist

Trident Services (Data Society Group)

Led the development of a No-Code ML Platform with integrated AutoML. Automated ML pipelines, achieving a 70% reduction in development time. Fine-tuned NLP models (BERT) for sentiment analysis achieving a 20% accuracy boost.

Jul 2021 - Jul 2022

Assistant System Engineer

Tata Consultancy Services (TCS)

Managed large-scale SAP HANA databases, reducing data downtime by 20%. Developed Tableau BI dashboards tracked monthly incidents and improved SLA by 30%.

Technical Toolkit

Core AI & ML

Generative AI Agentic AI NLP Computer Vision Deep Learning Machine Learning Ensemble Methods Time Series

Frameworks & Models

Transformers BERT Langchain LangGraph CrewAI Mistral / LLMs PySpark

Infrastructure & MLOps

AWS (Sagemaker) Azure GCP (Vertex AI) MLFlow DVC Docker Kubernetes (K8s) Databases (MySQL, PostgreSQL, MongoDB, DynamoDB) Vector DBs (Chroma, Pinecone, Milvus, Qdrant, Azure AI Search, AWS OpenSearch)

Programming & BI

Python R C SQL Tableau Power BI

Featured Projects & Open Source

Agentic Text-to-SQL Chatbot

Text-to-SQL Chatbot Architecture
Watch Demo

Architected an agentic Text-to-SQL chatbot utilizing stateful LangGraph workflows to translate natural language user queries into highly optimized SQL commands. Deployed on an AWS EC2 instance with a custom conversational UI, integrating schema pruning and execution validation for secure, real-time database querying against a sales dataset.

LangGraph SQL / AWS EC2 Agentic AI

Agentic RAG Chatbot: MetaMech

Chatbot Architecture
Watch Demo

Architected an enterprise-grade Agentic RAG system on the Azure ecosystem, utilizing Azure OpenAI for intent routing and generation. Orchestrated LangChain agents with Azure AI Search for high-dimensional semantic retrieval, incorporating multi-turn memory management to deliver highly accurate, context-aware responses.

Langchain Azure AI Search Azure OpenAI

No-Code ML Platform

No-Code ML Architecture
Watch Demo

Designed a modular, scalable No-Code ML Platform with an AutoML core. Hosted on AWS microservices, the backend intelligently handles missing data imputation, hyperparameter tuning via Bayesian optimization, and model selection, seamlessly exposing operations via REST APIs to a dynamic drag-and-drop frontend.

AutoML Python AWS (Sagemaker)

Multi-AI Agent Travel Assistant

Multi-AI Agent Architecture
Watch Demo

Engineered a distributed multi-agent system using CrewAI and LangGraph to model complex workflows. Implemented specialized autonomous agents with stateful graph-based execution, dynamic tool calling for real-time pricing data, and iterative consensus mechanisms to generate highly optimized itineraries at scale.

CrewAI LangGraph Python

AutoNPI ML Engine

AutoNPI ML Architecture
Watch Demo

Developed a robust ML intelligence layer for New Product Introduction (NPI), integrating Scikit-Learn predictive models into complex SQL data pipelines. Deployed via a continuous integration MLOps framework, leveraging historical feature sets to predict manufacturing bottlenecks and supply chain risks.

Scikit-Learn MLOps SQL

Voice-Driven Pizza Bot

Pizza Bot Architecture
Watch Demo

Built a voice-first conversational agent integrating real-time speech-to-text (STT) and text-to-speech (TTS) pipelines with a fine-tuned Mistral LLM backend. The system employs zero-shot entity extraction and robust dialogue management to handle unconstrained queries and multi-intent slot filling under noisy audio conditions.

Mistral Audio Processing NLP

NLP Text Summarizer (Pegasus)

NLP Summarizer Architecture
Watch Demo

Implemented an abstractive NLP pipeline utilizing the Pegasus transformer model, optimized for document-level understanding. Leveraged deep learning techniques including dynamic sequence padding and custom tokenization, evaluated via BERTScore and ROUGE to ensure semantic fidelity of the generated summaries.

Transformers BERT/Pegasus Deep Learning

Compensation Fitment System

Compensation Fitment Architecture
Watch Demo

Developed a predictive compensation analytics engine utilizing optimized Random Forest regressors. Ensured model transparency by integrating SHAP values to interpret global/local feature contributions, combined with a built-in rule-engine triggering a human-in-the-loop email escalation workflow for TA-CHRO.

Random Forest SHAP Classical ML

Fitment MCP Server (Voice)

MCP Server Architecture
Watch Demo

Architected a scalable Model Context Protocol (MCP) server bridging Voice AI with the Compensation Fitment pipeline. Designed a low-latency stateful server handling audio streams, transcribing intents, and querying classical models seamlessly via structured JSON-RPC communication secured by a relational database backend.

MCP Server Voice AI Databases

Community & Playlists

I regularly host community sessions and have created comprehensive playlists on modern data science tools including LangGraph, Tableau, and Power BI.

×