Building products at the intersection of AI × enterprise complexity
6+ years building and delivering enterprise B2B technology, with hands-on ownership of AI products that automate complex support workflows at scale.
Selected work & outcomes.
Converts support troubleshooting logic into AI-generated responses, combining error classification, customer context and resolution guidance.
PROJECT 02Retrieves relevant historical troubleshooting patterns and surfaces citation-backed resolution recommendations directly within the engineer workflow.
PROJECT 03Identifies tickets that can be safely merged or linked using precision-first "golden match" criteria designed to reduce duplicate analysis while protecting against false merges.
Projects
Intelligent First Response
AI-generated first responses and customer handling notes for support tickets
• Problem: Engineers spent time on the first read of every ticket, decoding the error and hunting SharePoint for customer handling instructions.
• Action: Pitched surfacing handling notes in Zendesk, defined 18 error categories and the AI response format, and worked with engineering on the AWS architecture.
• Result: Live since March 2026, with ~150K tickets processed, a 97.1% response rate, and an 89.7% engineer acceptance rate.
Stack: Python, AWS (API Gateway, Lambda, DynamoDB, S3, SQS, CloudWatch, Bedrock, EventBridge, CodePipeline, CodeBuild, ECS Fargate), Zendesk, LangChain, Blue-Green deployment, Snyk, HMAC-SHA256
Past Ticket Analysis
RAG over resolved tickets, so analysis reflects how engineers actually solved similar issues
• Problem: Generic AI first responses didn't reflect how engineers had actually solved similar tickets before.
• Action: Analyzed 22K+ resolved tickets into reusable "signatures" and built a RAG-powered Zendesk side-app with cited similar tickets and a chat interface.
• Result: Live since July 2026, with ~58K tickets processed, a 98.6% response rate, and an 80.6% acceptance rate, and signatures consolidated from 503 to 374 to push that higher.
Stack: Python, Claude Sonnet/Haiku 4.5, Titan Text Embeddings v2, OpenSearch (Vector Store), AWS serverless stack, Zendesk App Framework, LangChain, FastAPI
Ticket Merge/Link Agent
Automating duplicate ticket merging to cut wasted compute and engineer effort
• Problem: About 48% of tickets are manually merged because they share a root cause, wasting engineer effort and LLM tokens.
• Action: Spent 3 months working with engineers across 207 customers to write a 7-revision SOP, then designed an agent that merges only on an exact "golden match".
• Result: Targeting automation of 80% of manual merges, currently in testing.
Stack: Python, AWS (API Gateway, Lambda, DynamoDB, S3, SQS, CloudWatch, Bedrock, EventBridge, CodePipeline, CodeBuild, ECS Fargate), Zendesk, LangChain, Blue-Green deployment
Auto-Resolution of Defined Ticket Types
Exploring autonomous resolution for well-understood ticket categories
• Problem: Some well-understood ticket types still need engineer time even though the fix is predictable.
• Action: Helping explore which ticket types could safely be resolved without engineer intervention, building on the trust earned by Intelligent First Response.
• Result: Still in the discussion phase with no timeline, so I'd label it "Exploring" on the site.
Qualification
Work Experience
Senior Engineer, B2B Supply Chain
Integration Specialist, B2B Supply Chain (SPS Commerce)
Education
College
Joint Secretary, Association of Students of Electrical Engineering (ASEE) [2018-19]
• Overall Head of the Association: Organized the Inter-College Engineering Project Exhibition, Reboundz 2018 (inter-college basketball tournament), and Bharatiyam 2019 (inter-college cultural fest). Earlier, as Marketing Associate, raised funds for Reboundz 2017.
Marketing Manager, Entrepreneurship Development Cell (E-Cell) [2016-18]
• Marketing Manager (2018): Raised funds with the team for e-Summit 2018, which invited nominations from colleges and startup founders for talks.
• Senior Associate (2017-18): Helped organize Start-up Conclave 2017, a job fair connecting Pune startup heads and students for internships and full-time roles. I secured a Digital Marketing internship through this event.
• Junior Associate (2016-17): Helped organize e-Summit 2016 and the Social Entrepreneurship Challenge 2017, a hackathon on solutions to social problems. I presented a solution giving scrap dealers a place in the refurbished-products market.
School
[Madhya Pradesh Board of Secondary Education, Madhya Pradesh]
Intermediate - 12th 2015 - 2016 Score :- 83.40%Scholars' Den School, Khandwa (Madhya Pradesh)
[Madhya Pradesh Board of Secondary Education, Madhya Pradesh]
Matriculation - 10th 2013 - 2014 Score :- 85.17%