Hello World, I'm

Avinash Mishra

Manager · Data ScientistGenAI & LLM SpecialistAI Agent Systems Architect

Experienced AI leader building enterprise-scale GenAI products, RAG pipelines & AI-Agent systems. — Driving enterprise AI transformation.

17+
Team Members Led
12+
Years Experience
Avinash Mishra
ETHR Award 2025
BFSI Exceller 2025
Oracle GenAI Certified

Building AI That Matters

I'm a seasoned AI leader with 13+ years of experience transforming data into enterprise intelligence. Currently managing the Analytics & AI function at State Bank of India — India's largest public sector bank.

My expertise spans GenAI product development, RAG pipelines, LLM-powered agent systems, and responsible AI governance. I've led AI initiatives generating billions in business value, pioneering enterprise chatbots, regulatory AI, and hyper-personalization engines.

As founder of Data-Science-Trend and author of an open-source Data Science Essentials iPy-Book, I'm committed to democratizing AI knowledge globally.

Mumbai, India Location
Manager, Data Science @ SBI Current Role
English · Hindi · Japanese (N4) Languages
🤖

GenAI Pioneer

Built award-winning askSBI chatbot — India's first enterprise-scale AI Agent at SBI, serving 250K internal users.

📊

Scale Expert

Delivered enterprise-scale ML solutions with models generating significant measurable business value across banking operations.

🏛️

AI Governance Leader

Designed Responsible AI Framework (R-AI) for safe LLM deployment in India's largest regulated banking environment.

🎯

Team Builder

Acquired ¥25M JPY budget to build AI team from scratch in Japan.

Work Experience

Manager (Data Science)

State Bank of India · India
Mar 2021 – Present

Leading AI Centre of Excellence, transforming SBI's analytics department into an enterprise GenAI powerhouse.

  • askSBI GenAI Chatbot: 250K users · ETHR Award 2025 · BFSI Exceller Award 2025
  • CDNA-CLV: 360° customer persona engine · Omnichannel hyper-personalization at enterprise scale
  • Pre-approved Traders Loan: ML model targeting savings account holders with UPI patterns for SME lending
  • Designed Responsible AI (R-AI) governance framework for LLMs in banking
  • IIT-Bombay Technical Committee member for SBI DS Foundation Hub
  • 21+ senior recognitions from DMDs, CGMs, GMs for AI/DS excellence
GenAILLMsRAGPythonAzureKubernetesLangChain

Data Scientist

Rakuten · Japan
Dec 2020 – Feb 2021

Senior member of the Insight Analytics Group. Built predictive models for e-commerce delivery optimization.

  • At-home prediction model for Rakuten Ichiba last-mile delivery optimization
  • Built multi-point data fetcher package for seamless Rakuten data source integration
CatBoostPythonJupyterPrediction Modeling

Data Scientist

Catalina Marketing Japan (CMJ) · Japan
Jun 2018 – Nov 2020

Led multinational Data Science team delivering AI/ML solutions for retail and CPG clients. Managed cross-country collaborations across US, France, and Germany.

  • Media Mix Tracker: OLS + LightGBM/NN blend with SHAP explanations for Mizkan & Suntory
  • Likely Buyer model: Significant lift in campaign targeting accuracy for marketing campaigns
  • Customer Segmentation: Automated pipeline for 90M monthly transactions
  • New Product Launcher: ML forecaster + Dash dashboard for Suntory & Somi MFG
  • Retail_AutoML: Industry-specific autoML platform for CPG regression & classification
PythonSparkLightGBMword2vecDashPlotlyScala

Software Engineer (Data Science Lead)

Paradigmshift.io · Japan
Sep 2016 – May 2018

Led multinational team building state-of-the-art NLP systems. Reported to COO. Acquired ¥25M JPY budget to expand AI capabilities from scratch.

  • RepChecker: NLP tool for hotel OTA review analysis supporting 5 languages
  • Price RecSys: Recommendation engine for 100% room occupancy optimization
  • Smart web crawler: Distributed architecture with Cassandra cluster storage
  • Acquired ¥25M JPY to build and train an AI team from scratch
NLPPythonCassandraKafkaAirflowAWSMicroservices

Software Engineer

Rakuten · Japan
Oct 2014 – Sep 2016

Worked in Global Search Platform (GSP) team for Rakuten Ichiba. Built ML models for card lead generation and NLP-based transaction categorization.

  • XGBoost model for Rakuten Pink card lead generation (Japanese women segment)
  • NLP pipeline for card transaction categorization (travel, groceries, utilities, etc.)
XGBoostNLPJavaSolrElasticSearchLuceneCassandra

Systems Engineer

Infosys Ltd. · India
Mar 2011 – Aug 2012

Backend engineering for healthcare data management systems and employee portal development.

  • Molina Healthcare: SQL Server database with stored procedures, triggers, and functions
  • Employee SWAP portal: Web-based transfer management system
JavaSpringSQL ServerMySQLJSPStored Procedures

High-Impact Projects

🤖 GenAI Agent

askSBI GenAI Chatbot

Award-winning enterprise AI Agent system for India's largest bank. Built with RAG-based retrieval, hybrid cloud (Azure + Private Cloud), agent orchestration, and enterprise-grade guardrails.

250K internal users · ETHR Award 2025 · BFSI Exceller Award 2025
RAGLangChainAzureChromaDBLLMsFastAPIKubernetes
🧬 Hyper-Personalization

Customer-DNA & CLV Modeling

360° customer persona engine for 24K+ branches. Powers omnichannel hyper-personalization at enterprise scale using advanced clustering and RFM techniques.

Enterprise-scale deployment · 24K+ branches served
PythonXGBoostClusteringRFMSparkMLflow
💰 Business Growth ML

Pre-approved Traders Loan

ML product targeting savings account holders with UPI transactions to offer SME business loans. Uses Cross-Tab, RFM, clustering and RSM techniques for precision targeting.

Deployed in production FY23 · High-precision targeting for SME segment
PythonClusteringRFMCross-TabRSMSpark
⚖️ Regulatory AI

AI Governance Suite (Regulatory GPT)

Suite of AI Agents: Regulatory GPT, SOP-GPT, Deceased Account Settlement Bot, Project Finance Knowledge Bot, and Contact Centre Assistant — all with responsible AI compliance.

Enterprise-wide · Hallucination guardrails · Citation enforcement
LLMsRAGAgent OrchestrationLlamaIndexHaystackR-AI Framework
📈 Marketing Science

Media Mix Tracker (MMT)

Statistical analysis tool estimating marketing tactics' impact on sales. OLS for descriptive analysis, LightGBM/NN blend for prediction, with SHAP explanations for interpretability.

Piloted for Mizkan & Suntory Japan · SHAP-driven explainability
OLSLightGBMNeural NetworkSHAPPythonPlotly
🛒 Customer Intelligence

Customer Segmentation Pipeline

Automated pipeline segmenting 90 Million monthly transactions based on shopping behaviors. Powers personalized marketing campaigns. Includes Likely Buyer model for CPG clients.

90M monthly transactions · ~17% higher conversions with Likely Buyer model
PySparkK-meansword2vecHadoopScalaPython

Skills & Technologies

🤖

GenAI & LLM Frameworks

LangChainLlamaIndexHaystackSmolAgentsAutoGenCrewAIOpenAI GPTGeminivLLMOllamaHuggingFace Transformers
🕵️

AI Agent Systems

MCPA2A ProtocolFunction CallingPrompt OrchestrationVLMsOCR (Tesseract, LayoutLM, Donut)Multimodal AI
🔍

RAG & Vector Search

ChromaDBWeaviateFAISSCosmosDBElasticSearchSolrApache Lucene
🧪

Data Science & ML

PythonScikit-LearnTensorFlowPyTorchKerasXGBoostLightGBMMLflowPyCaretProphetSpark MLlibNLP / NLTK
☁️

Cloud & DevSecOps

Azure CloudAWSGCPDigitalOceanKubernetesDockerJenkinsCI/CDAirflowKafka
🗄️

Big Data & Databases

Apache SparkHadoopHiveKafkaCassandraMongoDBMySQLRedisOracle DBIBM DB2Netezza
📊

Visualization & BI

TableauDOMOMatplotlibKibanaDash by PlotlySeaborn

Languages & Web Frameworks

PythonSQLScalaRJavaJavaScriptBashPHPFastAPIFlaskDjangoVueJSNode.js

Awards & Achievements

Education & Learning

🎓

Master of Technology (M.Tech)

VJTI Mumbai — Software Engineering
2012 – 2014

Teaching Assistant: Data Structures & Algorithms, Software Engineering

🎓

Bachelor of Engineering (B.E.)

RGPV Bhopal — Computer Science & Engineering
2006 – 2010

Languages

🇬🇧
EnglishBusiness Proficiency
🇮🇳
HindiNative
🇯🇵
JapaneseN4 · Basic Conversational