Debasish Maji

Senior Software Engineer

Bengaluru, Karnataka, India10 yrs 3 mos experience
AI EnabledHighly Stable

Key Highlights

  • Expert in architecting reliable AI systems.
  • Led significant AI projects at Microsoft and Atlassian.
  • Proven track record in enhancing system scalability and performance.
Stackforce AI infers this person is a SaaS and Fintech expert specializing in AI systems architecture and scalability.

Contact

Skills

Core Skills

Ai EngineeringScalable System DesignAi PersonalizationScalability OptimizationPlatform DevelopmentMicroservices ArchitectureE-commerce DevelopmentApi Integration

Other Skills

AI AgentLLM AgentLLM EvaluationKnowledge EngineeringDeep LeadningLarge Language Model Operations (LLMOps)Large Language ModelsVector SearchDeep LearningReal-time ProcessingUsage MeteringEmerging TechnologiesContainer OrchestrationJavaREST

About

Architecting the reliability layer for modern AI systems. Currently working as a Senior AI Engineer at Microsoft, building enterprise-scale AI systems across Agentic AI, RAG, multi-agent orchestration, evaluation, grounding, reasoning, and production AI infrastructure. My work revolves around a problem I believe will define the next era of AI: How do we build AI systems that are not just intelligent, but reliable, observable, scalable, and trustworthy in real-world production? Over the years, I’ve worked across Microsoft, Atlassian, and PhonePe building large-scale AI/ML systems, retrieval architectures, distributed platforms, personalization systems, and enterprise infrastructure operating at massive scale. A major part of my recent work focuses on enterprise AI reliability: • AI evaluation systems • groundedness validation • synthetic evaluation pipelines • context engineering & retrieval quality • orchestration and agent reliability • scalable AI infrastructure for production workloads I’m particularly interested in the architectural shift happening across the industry: From standalone models → orchestrated AI systems. I believe the future of AI will not be defined by model size alone, but by the systems built around models - retrieval, orchestration, evaluation, memory, observability, and reliability. The hardest problems in AI begin after the demo works. Beyond building systems, I actively share practical insights on enterprise AI architecture, Agentic AI systems, evaluation, and reliable AI engineering - helping engineers move from experimentation to real-world production systems.

Experience

10 yrs 3 mos
Total Experience
3 yrs 1 mo
Average Tenure
11 mos
Current Experience

Microsoft

Senior Software Engineer

Jul 2025Present · 11 mos · Bengaluru, Karnataka, India · Hybrid

  • Thrilled to begin a new chapter at Microsoft, where I’m leading a team focused on building production-grade, real-world AI agent systems with measurable impact.
  • My work sits at the intersection of Agentic AI, LLM Evaluation, Knowledge Engineering, and Scalable System Design — with a strong emphasis on reliability, grounding, and enterprise readiness, not demos.
  • We go beyond “basic GenAI” by:
  • Advancing rigorous LLM & agent evaluation
  • Designing custom, tenant-aware knowledge bases
  • Delivering context-driven, intelligent AI agents that solve real business problems at scale
  • Energized to lead this mission and collaborate with exceptional engineers who believe in shaping the future of responsible, reliable AI.
  • Key Contributions & Impact
  • Led tenant-level customization leveraging advanced context engineering to deliver top-tier, high-precision AI agent quality in production.
  • Owned technical direction and delivery for multiple AI agent initiatives, driving architecture, execution, and production rollout.
  • Designed and implemented a scalable multi-agent architecture, enabling secure Agent-to-Agent (A2A) coordination using MCP-based communication for enterprise AI workloads.
  • Led end-to-end Helix workstreams for Workforce Insights, including contracts, grounding strategy, migration planning, and staged production deployments.
  • Drove context engineering and retrieval improvements, significantly improving grounding quality, tenant relevance, and retrieval effectiveness for production agents.
  • Recognized as IDC Give Champ for strong ownership, cross-team impact, and org-level knowledge sharing.
  • Led organization-wide knowledge sessions on RAG, hybrid search, LLM evaluation, and agent design patterns.
  • #Microsoft #AgenticAI #LLMEvaluation #AIEngineering #Leadership #EnterpriseAI #RAG #Helix #LifeAtMicrosoft
AI AgentLLM AgentAI EngineeringScalable System Design

Atlassian

Senior Engineering Associate

Aug 2020Jun 2025 · 4 yrs 10 mos · Bengaluru, Karnataka, India

  • User Embedding Service – Leading the development of a high-dimensional user embedding service leveraging large language models (LLMs), vector search, and deep learning techniques to significantly enhance AI-powered search ranking, personalization, and intelligent automation across Atlassian products. Currently collaborating cross-functionally to deliver impactful improvements in product relevance and user engagement.
  • Usage Tracking Service (UTS) at Scale – Actively designing and optimizing the Usage Tracking Service (UTS) to handle significant scalability demands. Implementing improvements to efficiently manage high traffic volumes, ensuring robust real-time AI feature integration, accurate usage metering, and reliable entitlement management for quota-based billing across Atlassian products.
  • Commerce Cloud Platform (CCP) Development:
  • Built Atlassian's CCP, leading to an increase in sales and revenue growth with a potential 10B Sales.
  • Enabled flexible offerings and promotions.
  • Led a huge number of EPIC, SPIKE, HLD, LLD, DACI and implementation
  • Secured our resources with fine grained access control using RBAC, Drove the initiative of resource authorisation and adoption across multiple services
  • Migration from Monolith to Microservices:
  • Led migration, achieving a 20% improvement in system agility and a 25% increase in scalability.
  • Implemented workflow management for cohort-based migrations.
  • Architectural Leadership:
  • Significantly contributed to a microservices-based architecture.
  • Achieved a 30% increase in system reliability and 20% improvement in scalability.
  • Performance Optimization:
  • Resolved critical bottlenecks, reducing response times by 90%. Also, We have achieved 99.99% SLO from 99.95%
  • Enhanced user experience with a 25% improvement.
  • RBAC Implementation:
  • Upgraded tech dependencies, including RBAC authentication.
  • Ensured a 15% enhancement in security and compliance.
Deep LeadningLarge Language Model Operations (LLMOps)AI PersonalizationScalability Optimization

Phonepe

Software Engineer

Jul 2018Aug 2020 · 2 yrs 1 mo · Bangalore

  • MVAS Platform:
  • Built the MVAS platform for PhonePe from the ground up, serving as the foundation for various merchant offerings (Merchant Bookkeeping, Cash In Cash Out, CRM), resulting in a ~30% increase in merchant traction.
  • Cross-Service Collaboration:
  • Led discussions with multiple Proof of Concept (POC) teams across different services.
  • Drove necessary changes in the new system and coordinated with upstream and downstream services, including Payment Service, Identity, and Merchant services.
  • Development of Khata and CICO Services:
  • Developed Khata (Bookkeeping) and CICO (Cash In Cash Out) services.
  • Utilized technologies such as Java Dropwizard, MySQL, REST, Aerospike, RabbitMq.
  • Implemented crucial techniques including Sharding Strategy, Async processing, Report for failure cases, Finite State Machine, and Monitoring & Observability using Grafana and Prometheus.
  • TOA Processing Automation:
  • Automated TOA (Token of Apology) for deemed failure transaction processing for merchants, leveraging the company-wide TOA process.
  • User Personalization Feature:
  • Developed a user personalization feature supporting suggestion-based functionalities like Recent Scans.
  • Implemented a data processing pipeline using Apache Spark, Airflow, Hive, and HBase.
  • Executed ETL processes to handle data, enabling high-performance API with latency less than 50 ms.
Emerging TechnologiesContainer OrchestrationPlatform DevelopmentMicroservices Architecture

Zopper.com

2 roles

Software Engineer

Jun 2016Jul 2018 · 2 yrs 1 mo

  • Implemented backward compatible versioned APIs to ensure smooth app upgrades with minimal impact for customers. Reduced customer churn rate by ~20%.
  • Implemented multiple online payment methods in addition to online cash modes which lead to increase in GMV from $15M to $25M.
  • Reduced home and product page load time by 30% and 35% respectively by introducing caching layers for aggregated data.
  • Built user accounts and payments platforms for Zopper’s hyperlocal e-commerce offering
  • Designed and developed backend online shopping functionality including order management, shopping carts, payments and user authentication modules
  • Helped materialize strategic financial partnerships through integration of 10+ payment/loan providers into the backend payments framework.
  • Lead a team with an initiative to containerise codebases using docker to increase efficiency and robustness of dev environments leading to a massive reduction in new environment setup time and effort.
  • Secured financial applications by implementing audit logs for liability protection, risk management and compliance.
  • Slashed support queries by over 50% by implementing refunds and transaction status checks using schedulers.
  • Improved online payment success rates by 25% through dynamic switching of gateways using custom rules and health checks.
  • Eliminated synchronisation delays for users by authoring an open source, multi-threaded, scalable SQS queue consumer package (github.com/arijeetmkh/hephaestus) leading to data ingestion throughput increase by 200%.
  • Setup the backend event streaming, consumption and broadcast layers using AWS SNS and SQS. Achieved a real-time target of upto 2.5 seconds from invoice creation to its reflection into respective systems.
Container OrchestrationREST APIsE-commerce DevelopmentAPI Integration

Software Engineer Intern

Jan 2016May 2016 · 4 mos

  • Have learned various new technologies : python, flask, django. And worked on a number of projects related to price syncing pipeline, Order Management System, Delivery Management System.
OptimizationProblem Solving

Education

Banaras Hindu University, Banaras

Master of Computer Applications (MCA) — Computer Science

Jan 2013Jan 2016

Midnapore College

Bachelor of Science (B.Sc.) — Mathematics Hons

Jan 2010Jan 2013

West Bengal Council of Higher Secondary Education

Science stream

Jan 2008Jan 2010

West Bengal Board of Secondary Education (WBBSE)

SSC

Jan 2007Jan 2008

Stackforce found 100+ more professionals with Ai Engineering & Scalable System Design

Explore similar profiles based on matching skills and experience