Software Engineering
System design, scalable architecture, clean code, testing, security, observability, reliability, performance, and operational ownership.
Software Engineer II at Microsoft | Copilot Agents & Platform
Sunil Gudivada.
I build production-ready AI agents and scalable platforms that turn complex ideas into reliable customer experiences.
Agent Pulse
AI Agents
Microsoft Copilot
Multi-Agent Systems
Agentic AI
Agent Orchestration
GraphRAG
TinyTroupe
Digital Twins
Persona Evaluation
Retrieval-Augmented Generation
LLM Evaluation
Prompt & Instruction Engineering
Model Context Protocol
Agentic Memory
Deep Research
SEVAL
2017
Amdocs
Intern
2018 - 2020
Oracle Financial Services
Software Engineer
2020 - 2021
Talentica Software
Full-stack product engineering
2021
WillHire
Software Engineer
2022 - Present
Microsoft
Software Engineer II
01 About
Engineering ambitious ideas into dependable products.
10+
high-impact AI, platform, and product initiatives delivered
-
2024
Viva Learning Excellence Award Engineering excellence and customer impact
-
2025
Trailblazer Award Innovation and high-impact delivery
-
2026
Trailblazer Award Recognized in consecutive years
-
4x
Top-Three Hackathon Finishes Within the Viva Learning team
-
Pinnacle
People's Choice Award Peer-recognized innovation
-
Build
Agent Platforms Micro Agents, Agent Pulse, and SEVAL quality tooling
-
Explore
Applied AI Innovation Teams Meeting Digital Twin, Persona Research, and Robo Rumble
-
Create
Independent & Social Impact System Design Studio and VoteMate startup enablement
Software Engineer II at Microsoft and senior AI systems developer delivering production-ready AI solutions. Led the creation of multi-agent capabilities and built 3 evaluation tools covering 700+ queries, accelerating feature onboarding and reducing defects. One patent filing is under review, with multiple Trailblazer Awards. Seeking to apply expertise in LLM evaluation, secure observability, and system design to build impactful AI products.
Applied AI Engineering
Agent orchestration, RAG and GraphRAG, LLM evaluation, grounding, persona simulation, experimentation, regression analysis, and agent observability.
AI platform architecture
Agent quality & evaluation
Product innovation
Production engineering
Technical leadership
02 Experience
A career built around ownership, scale, and measurable change.
Jan 2022 - Present
Microsoft
Software Engineer II
Driving architecture and end-to-end delivery for Microsoft Learning Agent and Viva Learning, with a focus on scalable AI systems, release quality, and customer impact.
AI Innovation Spotlight
- 01 / Observe Agent Pulse Unified agent quality, reliability, and performance signals to expose regressions earlier, accelerate investigation, and strengthen release decisions.
- 02 / Simulate Teams Meeting Digital Twin Combined GraphRAG and TinyTroupe to simulate realistic personas and meeting interactions, enabling agent behavior research before customer validation.
- 03 / Evaluate Customer Delight Automation Evaluated persona-based customer journeys through repeatable agent assessments that reveal experience gaps and guide high-value product improvements.
- 04 / Orchestrate Robo Rumble Led agentic architecture among 300+ global participants, creating task-extraction, orchestrator, and feedback-loop patterns for AI robotics.
- 05 / Empower VoteMate Startup Enablement Supported an NGO-focused application through Microsoft's startup incubation ecosystem, developing a custom RAG agent that researches candidate information and translates public web data into accessible voter insights.
- Architected and led Phase 1 delivery of a reusable Micro Agents model, replacing monolithic agent patterns and enabling Coach, Teach, and Deep Learning teams to build independently on a shared platform.
- Migrated Learning Agent to the Helix runtime and delivered Graph Connector, Microsoft 365 Search, AugLoop plugin, custom connector, and web chat integrations, expanding the agent to new content sources and licensed-user scenarios.
- Built the secure Loki-to-Viva Learning integration adopted by 5+ major features, reducing partner onboarding by 50% and debugging time by ~60%.
- Created 3 SEVAL quality tools for validation, analysis, and regression classification; established release gates across 700+ evaluation queries.
- Led 8 weekly beta-test cycles, identified 60+ defects before production, reviewed 100+ pull requests, and resolved URL-hallucination, CSP, and MSRC security blockers.
- Reduced the active incident backlog from 49 to 14, a 71% improvement, and lowered repeat escalations by ~40%.
- Shipped Copilot Academy, Learning Profile, One Shell, Fluent UI v9, provider management, and recent-learning experiences; addressed 100% of security recommendations for Learning Profile.
- Raised engineering capability through 30+ architecture forums, mentored engineers, progressed a patent filing (in progress), and translated multi-agent research into product roadmap proposals.
Apr - Dec 2021
WillHire
Software Engineer
Delivered sourcing, job distribution, search, and cloud infrastructure improvements for a recruiting technology platform.
- Reduced robotic sourcing cost by 40% by reusing resumes already available in the system.
- Improved job feed load time by 95% using Elasticsearch and Amazon S3.
- Improved applicant performance by 32% through third-party job distribution integrations.
- Built an in-house URL shortener with CloudFront, Lambda, and API Gateway.
Mar 2020 - Mar 2021
Talentica Software
Software Engineer
Built financial integrations, a configurable chatbot framework, and a backend-driven e-learning platform across Java, Python, and React.
- Reduced REST API response time by 70% by improving Spring caching.
- Created a configurable chatbot framework using Python, Flask, AIML, NLTK, and scikit-learn.
- Built a JSON-defined UI architecture for instant application updates.
- Developed a utility to compare data across MySQL and Apache Solr.
Jul 2018 - Mar 2020
Oracle Financial Services Software
Associate Application Developer
Developed microservices, automation, observability, and orchestration for financial crime and compliance products.
- Reduced provisioning installation from 15 minutes to 2 minutes through shell automation.
- Built Spring Boot microservices with Liquibase, Swagger, Gradle, and Jenkins.
- Added real-time monitoring with Grafana and Prometheus.
- Developed product orchestration using Docker and Kubernetes.
May - Jul 2017
Amdocs
Summer Intern
Built a Cortana and Arduino Windows application supporting voice, text, organizational information retrieval, and IoT device control.
03 Selected work
Systems designed to create leverage beyond a single feature.
Micro Agents for Learning
A reusable architectural pattern that decomposes a monolithic learning agent into focused capabilities, enabling Coach, Teach, and Deep Learning experiences while improving scalability and team autonomy.
SEVAL Quality Ecosystem
Validator, analyzer, and regression-classifier tools that moved agent quality earlier in the lifecycle and enabled data-driven release decisions across more than 700 evaluation queries.
Loki x Viva Learning
Secure, traceable platform foundations used by more than five major features. The integration moved partner onboarding from weeks to days and reduced debugging time by approximately 60%.
Copilot Academy & Agent Prototypes
General-availability learning experiences alongside explorations in orchestration, memory, deep research, adaptive learning cards, and MCP. Hackathon work produced nine ideas, five prototypes, and two Pinnacle selections.
Agent Pulse
An AI-agent observability and quality platform that brings evaluation results, reliability signals, and performance trends into one view. Agent Pulse helps engineering teams identify regressions earlier, diagnose behavior changes faster, and make confident, data-informed release decisions.
Teams Meeting Digital Twin
A GraphRAG-powered digital twin of Teams meetings built with TinyTroupe. The prototype models distinct personas, organizational context, and realistic meeting interactions to simulate how people and AI agents collaborate before conducting customer-facing validation.
Persona Research & Customer Delight
A persona-driven research and automation framework for evaluating AI agents against realistic customer goals, behaviors, and expectations. It converts customer scenarios into repeatable evaluations that surface experience gaps and help teams prioritize improvements with the greatest customer value.
System Design Studio
A visual learning studio for exploring system design concepts and architecture solutions. The product helps engineers reason about scalability, distributed systems, and design trade-offs through a focused, interactive experience.
Conceived before the mainstream GPT era; later evolved with AI-assisted curation to organize and enrich its system-design problem library.
Explore the live studioRobo Rumble
Led core agent architecture work among 300+ global participants, developing task-extraction, orchestrator, and feedback-loop patterns for Raspberry Pi and AI HAT-based robotics. Leadership demonstrations helped advance Project Cosmos and Mini Andro humanoid-robot integration.
VoteMate
Supported an NGO-focused civic application through Microsoft's startup incubation ecosystem. Developed a custom RAG-powered AI agent that combines candidate information with relevant public web research and presents clear, accessible insights to help voters make more informed decisions.
04 Recognition
Consistent impact, recognized by peers and leadership.
Two-time Trailblazer Award
Received the Trailblazer Award in consecutive years, recognizing sustained innovation, technical leadership, and high-impact delivery at Microsoft.
Viva Learning Excellence Award
Recognized for engineering excellence and meaningful contributions that advanced the quality and customer experience of Viva Learning.
Four Top-Three Hackathon Finishes
Placed among the top three in four hackathons within the Viva Learning team, including Robo Rumble, demonstrating a consistent ability to turn ambitious AI ideas into compelling, working prototypes.
People's Choice Award
Won the People's Choice Award at Pinnacle, reflecting strong peer support and the ability to communicate an innovative technical idea with broad organizational appeal.
05 Capabilities
Deep enough for architecture. Broad enough to ship.
AI Systems & Agent Platforms
Designing production-oriented agent platforms that combine architecture, orchestration, grounding, simulation, observability, and measurable quality.
- Multi-Agent Architecture
- Agent Orchestration
- GraphRAG & RAG
- LLM Evaluation
- Agent Observability
- Persona Simulation
- Agentic Memory
- MCP
Application Engineering
Java, Spring Boot, Spring Cloud, Python, Flask, React, JavaScript, REST, GraphQL
Cloud & Platforms
Azure, AWS, App Engine, EC2, Lambda, S3, CloudFront, API Gateway, Kubernetes Engine
DevOps & Reliability
Azure DevOps, Docker, Kubernetes, Jenkins, Gradle, Maven, telemetry, incident response
Data & Search
MySQL, Oracle Database, Elasticsearch, Apache Solr, data integration, caching
Engineering Excellence
Building secure, observable, testable, and maintainable systems with disciplined architecture and complete operational ownership.
- System Design
- Reliability
- Security by Design
- Quality Gates
- Observability
- Performance
- Incident Ownership
- Mentorship
06 Education
The foundation behind a career of continuous learning.
Bachelor of Technology, Computer Science
National Institute of Technology, Rourkela
Student Placement Coordinator during the final year.
Start a conversation
Let's build what matters.
Interested in AI platforms, scalable product architecture, or ambitious engineering challenges? I would be glad to connect.