Debasish Maji - AI/ML Leader
World-Class AI/ML Leader & Innovator

Debasish MajiDebasish Maji

Pioneering the Future of Generative AI & LLMs

Leading AI Agent Development at Microsoft | Building RAG-based systems that transform enterprise workflows | 9+ years driving innovation at Microsoft, Atlassian, and PhonePe

Schedule a Call
Work With Me
$1B+
Revenue Impact
550M+
Users Served
20%
AI Accuracy Boost
40%
Time Reduction
Scroll to explore

Leadership in AI Innovation

Transforming vision into reality through technical excellence and strategic leadership

AI Innovation Pioneer

Leading AI Agent Development for Personal Insights at Microsoft, implementing cutting-edge Agent-to-Agent communication and MCP protocol for enterprise-grade AI systems.

Revolutionizing how enterprises leverage AI for intelligent insights

Performance Excellence

Architected Rovo Agent at Atlassian, a RAG-based AI system that improved accuracy by 20% and developed user embedding systems enhancing search relevance by 12% for 10M+ monthly active users.

Measurable improvements in product performance and user experience

Scale Master

Built data pipelines at PhonePe serving 550M+ users and processing 6.71B annual transactions with sub-50ms latency using sophisticated microservices architecture.

Enterprise-scale systems handling massive user bases

Business Impact Driver

Enabled $1B+ in revenue through scalable usage tracking and billing solutions. Reduced code review time by 40% with AI-powered automation, directly impacting developer productivity.

Tangible business outcomes and operational efficiency gains

Driving AI Innovation

Transforming complex AI challenges into scalable, production-ready solutions

Debasish Maji working on AI systems

Building the Next Generation of AI Systems

Leading AI Agent development with cutting-edge LLM architectures and RAG systems

Advanced AI/ML Engineering

9+ years of software engineering with 4.5+ years focused on AI/ML and 3+ years pioneering Generative AI solutions. Deep expertise in LLM fine-tuning, quantization, RAG, and scalable AI infrastructure.

Research & Innovation

Conducting cutting-edge research on hallucination reduction in LLMs through architectural enhancements, hybrid search (dense & sparse retrieval), and self-supervised learning to improve accuracy and trustworthiness.

12%
Query Relevance Boost
LLM-powered search optimizations at Atlassian
40%
PR Review Time Reduction
AI-assisted engineering automation
20%
Retrieval Accuracy Improvement
Rovo Agent RAG system innovation

Beyond Technology

Leadership and innovation thrive on diverse experiences and fresh perspectives

Exploring nature and seeking inspiration

Finding balance through exploration

Exploring nature and seeking inspiration

Life on the water

Embracing adventure and new perspectives

Life on the water

Coastal inspiration

Drawing creativity from diverse experiences

Coastal inspiration

Bringing a holistic approach to AI leadership — combining technical excellence with global perspectives, creative thinking, and a commitment to continuous learning through diverse life experiences.

Community Initiative

Learn AI by Building

A collaborative initiative where we learn AI together by building real-world projects. No lectures. No prerequisites. No fees.

Visit WebsiteRead LinkedIn Post

Over the past few months, many people have reached out - not just to talk about AI, but to truly understand it, to learn by building, and to explore how it can solve real-world problems.

That genuine curiosity has been inspiring. So I decided to spend some of my personal time on weekends working with you on small, real-world AI projects.

Nothing formal - just a collaborative approach to learn together. No prerequisites - just bring your curiosity and a mindset to learn by experimenting.

Working on AI/ML and agent-based projects at Microsoft, Atlassian, and PhonePe - combined with learning from Coursera and DeepLearning.AI, researching AI hallucinations, and building research agents - has shaped how I think about intelligent systems. I genuinely believe we can build something meaningful together.

Learn by Building

No lectures, no theory dumps. Just hands-on projects that solve real-world problems

Collaborative Learning

Work in teams, share knowledge, and grow together with a global community

100% Free

No fees, no prerequisites, no barriers. Just bring your curiosity and willingness to learn

Real Projects

Build production-grade AI systems - RAG, agents, LLMs, and intelligent applications

Join the Initiative

Don't worry about it getting crowded - we can run multiple projects in parallel and form teams based on time zones. Share with friends or colleagues who might enjoy building something hands-on together!

More Initiatives

Community & Education Impact

Training people in AI/ML and building products that make a difference in the community

Community & Education

Learn AI By Building

1000+ Learners

A collaborative community initiative where people learn AI/ML by building real-world projects together. No fees, no prerequisites - just hands-on learning with global teams.

Collaborative LearningWeekend SessionsReal Projects
Visit Project
Training & Workshops

Thrive With AI

Professional Training

Educational platform focused on helping professionals and teams thrive in the AI era. Practical training, workshops, and resources for AI transformation.

Corporate WorkshopsAI TransformationHands-on Training
Visit Project
Health Tech

Care Of Life - Mental Health AI Assistant

Mental Wellness

An empathetic AI assistant designed to provide mental health support, guidance, and resources. Built with responsible AI practices and user privacy at core.

Empathetic AIPrivacy-First24/7 Support
Visit Project

Weekend passion projects allow me to experiment with cutting-edge AI technologies, build products that matter, and share knowledge with the community. From mental health support to AI education, each project aims to make a positive impact.

Project Portfolio

Featured Projects

Production-scale AI systems driving measurable business impact

Microsoft

Workforce Insight Agent

AI-powered agent system providing intelligent insights for workforce management and optimization using LLMs and RAG architecture

15+

Data Sources

<2s

Query Speed

92%

Accuracy

Key Highlights

  • •Multi-source data integration and synthesis
  • •Natural language query processing
  • •Real-time workforce analytics and recommendations
GPT-4RAGAzure AISemantic KernelAutoGenAzure OpenAIVector DB

Microsoft

Agent Evaluation Framework

Comprehensive evaluation system for AI agents, measuring performance, reliability, and quality across multiple dimensions

1000+

Test Cases

20+

Metrics

95%

Coverage

Key Highlights

  • •Automated evaluation pipeline
  • •Multi-dimensional quality assessment
  • •Performance benchmarking and regression detection
PythonMLflowAzure MLCustom MetricsA/B Testing

Atlassian

Rovo Agent - RAG-based AI System

Advanced Retrieval-Augmented Generation (RAG) AI system providing real-time intelligent assistance with enhanced accuracy through fine-tuning and quantization

+20%

Accuracy Boost

Low

Latency

Reduced

Hallucinations

Key Highlights

  • •Optimized retrieval pipelines with FAISS and vector search
  • •Reduced LLM hallucinations through enhanced retrieval efficiency
  • •Integrated into Atlassian Intelligence ecosystem for automated insights
  • •Enabled intelligent decision-making and AI-driven workflow interactions
RAGFAISSVector SearchEmbeddingsLLM Quantization

Atlassian

AI-Powered Search in Atlassian

Next-generation intelligent search system leveraging LLMs and semantic understanding to improve content discovery across Atlassian products

550M+

Users Impacted

+12%

Relevance Boost

<200ms

Query Speed

Key Highlights

  • •Semantic search with intent understanding
  • •Cross-product search experience
  • •Personalized ranking and recommendations
LLMsVector SearchElasticsearchGraphQLReact

PhonePe

Real-Time Fraud Detection System

ML-powered fraud detection processing 100M+ transactions daily with 95% accuracy and sub-100ms latency

100M+

Transactions/Day

95%

Accuracy

<100ms

Latency

Key Highlights

  • •Real-time feature engineering pipeline
  • •Ensemble models with online learning
  • •Reduced false positives by 30%
TensorFlowKafkaSparkHiveReal-time ML

Professional Journey

9+ years of driving AI/ML innovation across industry-leading companies

Featured Role

Microsoft

Senior Software Engineer
Leading AI Agent Development for Personal Insights in PAX
07/2020 - Present
Bangalore, India
  • Leading a team to build an AI agent for Personal Insights within the PAX group
  • Implementing sophisticated Agent-to-Agent communication for seamless coordination
  • Leveraging MCP protocol for secure, standardized interactions across systems
  • Designing scalable Multi-Agent framework for contextual and adaptive insights
  • Ensuring enterprise integration with Microsoft's ecosystem for employee-facing use cases
Featured Role

Atlassian

Senior Engineering Associate
Developer - CCP, Central AI
08/2020 - 06/2025
Bangalore, India
  • Designed and deployed the Rovo Agent, a RAG-based AI system, improving accuracy by 20%
  • Developed user embedding system improving search relevance by 12% for 10M+ MAUs
  • Led development of AI-powered Code Review Agent, reducing review time by 40%
  • Architected real-time ETL-based usage tracking system achieving 10K RPS
  • Enabled $1B+ in revenue through scalable usage tracking and billing solutions

PhonePe

Software Engineer
Payment and Merchant Ecosystem & Data
08/2018 - 08/2020
Bangalore, India
  • Engineered data pipelines serving 550M+ users and processing 6.71B annual transactions
  • Developed near real-time recommendation systems with user embeddings (NCF + FAISS)
  • Built MVAS platform driving 27% increase in merchant traction
  • Achieved sub-50ms latency with robust microservices architecture

Zopper

Software Engineer
Hyperlocal E-commerce Solutions
08/2016 - 08/2018
Bangalore, India
  • Led development of user account management and payment processing modules
  • Scaled GMV from $15M to $25M through payment optimization
  • Integrated 10+ payment/loan providers with optimized gateway switching

Technical Mastery

Comprehensive expertise across AI/ML, engineering, and senior leadership

AI/ML & Data Engineering

LLMs (GPT, BERT, T5, Claude, Llama)
RAG Systems
Fine-tuning & PEFT
Quantization
Semantic Kernel
AutoGen
LangChain
LangGraph
Prompt Engineering
PyTorch
TensorFlow
Keras
Hugging Face
AWS SageMaker
Azure ML
FAISS
Pinecone
ChromaDB
Weaviate
Spark
Scikit-learn
XGBoost
Neural Networks
NLP & Transformers
Computer Vision
Reinforcement Learning
MLOps
Model Optimization
A/B Testing
Feature Engineering

Software Engineering

Java
Python
C++
JavaScript/TypeScript
Spring Boot
FastAPI
Flask
Django
Microservices
REST APIs
GraphQL
gRPC
System Design
Design Patterns
Distributed Systems
Event-Driven Architecture
Clean Code
TDD/BDD
Agile/Scrum

Cloud & DevOps

AWS (EC2, S3, Lambda, SageMaker)
Azure (OpenAI, ML Studio, Functions)
GCP
Kubernetes
Docker
Terraform
Helm
CI/CD (Jenkins, GitHub Actions)
GPU Acceleration (CUDA)
Monitoring & Observability
ELK Stack
Prometheus
Grafana
Infrastructure as Code

Data Processing & Storage

MongoDB
PostgreSQL
MySQL
Redis
Elasticsearch
Kafka
RabbitMQ
Hadoop
Spark
Airflow
Data Warehousing
ETL Pipelines
Stream Processing

Leadership & Management

Team Leadership
Technical Mentorship
Cross-functional Collaboration
Stakeholder Management
Product Strategy
Technical Roadmapping
Hiring & Talent Development
Performance Management
Agile Leadership
Conflict Resolution
Decision Making
Resource Planning

Strategic & Business Skills

AI Strategy & Vision
Business Impact Analysis
Cost Optimization
Scalability Planning
Risk Management
Vendor Evaluation
Technical Presentations
Executive Communication
Project Management
Budget Planning
KPI Definition
Change Management
Insights & Articles

AI Insights

Sharing knowledge and perspectives on AI/ML trends, techniques, and best practices

AI Trends
Trending

The Future of RAG Systems: Beyond Simple Vector Search

Exploring advanced techniques in retrieval-augmented generation including hybrid search, re-ranking, and context compression for production systems.

8 min read
Jan 15, 2024
MLOps
Trending

Deploying LLMs at Scale: Lessons from 550M+ Users

Deep dive into infrastructure patterns, cost optimization, and reliability engineering for large-scale LLM deployments in production.

12 min read
Jan 8, 2024
Technical Leadership

Building High-Performance AI Teams: A Practical Guide

Strategies for hiring, mentoring, and scaling AI engineering teams based on experience at Microsoft, Atlassian, and PhonePe.

6 min read
Dec 28, 2023
AI Engineering

Fine-tuning vs RAG vs Prompt Engineering: When to Use What

Comprehensive comparison of different LLM adaptation techniques with real-world examples and performance benchmarks.

10 min read
Dec 20, 2023

Education & Credentials

Strong academic foundation and continuous learning through advanced certifications

Master of Computer Applications

Computer Science
Banaras Hindu University (Tier 1)
Varanasi, India2013 - 2016

Bachelor of Science

Mathematics Honours, Mathematics And Statistics
Midnapore College
Midnapur, India2010 - 2013

Professional Certifications

Datathon Champion (Atlassian)
Deep Learning Specialization (DeepLearning.ai)
Machine Learning (Stanford | Online)
Data Manipulation at Scale: Systems and Algorithms (University Of Washington)
Open for Opportunities

Let's Work Together

I'm available for consulting, freelance projects, teaching, mentorship, speaking engagements, and collaboration opportunities.

Book a time to discuss how we can work together to bring your AI vision to life.

Consulting

AI/ML strategy, GenAI implementation, and LLM architecture consulting for your enterprise

Book a call

Freelance Projects

Custom RAG systems, AI agent development, and ML pipeline implementation

Book a call

Teaching & Workshops

Technical training on Generative AI, LLMs, and enterprise ML systems

Book a call

1:1 Mentorship

Career guidance, technical mentorship, and AI/ML expertise sharing

Book a call

Podcast & Talks

Speaking engagements on AI innovation, leadership, and technology trends

Book a call

Collaboration

Research partnerships, startup advisory, and strategic AI initiatives

Book a call
Schedule a 30-Min Call

Free initial consultation • Flexible scheduling • Global availability

Frequently Asked Questions

Questions About AI/ML Training

Everything you need to know about learning AI/ML and transitioning to an AI career

The best AI courses for working professionals are hands-on, project-based programs that fit around work schedules. Our Learn AI by Building program offers free weekend sessions where you build real-world Generative AI, LLM, and AI Agent projects with guidance from a Senior Microsoft AI Engineer. No prerequisites required - just bring your curiosity.

Software engineers can transition to AI/ML by: 1) Learning foundational concepts through hands-on projects, 2) Building a portfolio of AI projects like RAG systems, chatbots, and AI agents, 3) Getting mentorship from industry experts, 4) Leveraging your existing programming skills. Weekend AI courses designed for developers make this transition practical while working full-time.

Generative AI refers to AI systems that can create new content - text, code, images, and more. You should learn it because: 1) It's transforming every industry, 2) Companies are actively hiring for GenAI skills with average salaries of $150K+, 3) It increases productivity by 40-60%, 4) Demand for AI skills is growing exponentially. Learning through hands-on projects is the most effective approach.

Yes, 100% free! Learn AI by Building is a community initiative with no fees, no hidden costs, and no prerequisites. The program is taught by Debasish Maji, a Senior AI Engineer at Microsoft, who dedicates weekend time to help professionals learn AI by building real projects together. Join 500+ professionals already learning for free.

Unlike Coursera or Udemy which offer pre-recorded videos, our program provides: 1) Live instruction with real-time Q&A, 2) Hands-on project building, not just theory, 3) Direct mentorship from an industry expert at Microsoft, 4) A supportive community of 500+ professionals, 5) Weekend-friendly schedule for working professionals. You learn by building real AI systems.

With dedicated weekend learning, you can gain job-ready AI/ML skills in 3-6 months. The key is hands-on practice: Build 3-5 real projects (chatbots, RAG systems, AI agents), understand core concepts through implementation, and leverage your existing software engineering skills. Our bootcamp-style weekend learning accelerates this timeline significantly.

You'll learn the most in-demand AI technologies: Large Language Models (LLMs), ChatGPT/GPT-4 integration, Azure OpenAI, Semantic Kernel, AutoGen, LangChain, RAG (Retrieval Augmented Generation) systems, Vector Databases (FAISS, Pinecone, ChromaDB), AI Agents, and production ML deployment. All through hands-on projects.

Our AI/ML training is available globally with flexible weekend sessions. We have participants from the United States, United Kingdom, India, and many other countries. Sessions are designed to accommodate different time zones, and you can join from anywhere with internet access.

Still have questions?

Book a Free Consultation

Let's Connect

Interested in discussing AI/ML innovations, collaboration opportunities, or just want to connect?

Feel free to reach out through any of these channels.

Email
debasishmath92@gmail.com
Phone
+91 9198390240
LinkedIn
debasish-maji
GitHub
DebasishMaji
Based in Bangalore, India
Open for meaningful opportunities in UK and USA

© 2025 Debasish Maji. All rights reserved.

Chat on WhatsApp