Bachelor's in Electronics & Communication
GGSIIPU Delhi. Completed in 2020.
Senior AI Engineer & AI Product Manager
Sumneet Kaur
I find the problems worth solving with AI, then build and ship the systems that solve them.
I work on both sides of an AI product. With stakeholders, I find the workflows worth automating and agree what good output looks like. As an engineer, I build the agents, retrieval, evaluation and guardrails, and the data platform that makes their output safe to put in front of customers. At my last company that cut manual review by 60% and listing friction by 43%, across 7 automation systems in production.
Multi-Agent Systems
RAG & Evaluation
Guardrails & Provenance
AI Product Management
Roadmaps & Delivery
Cloud & Data Platforms
The journey so far
From electronics engineering to enterprise delivery, then to building and owning production AI as both engineer and product manager.
Technology Analyst, Infosys
Platform engineering, DevOps and delivery for enterprise clients, including a 1,200-repository migration.
Moved to Sydney
Started Master of Computer Science at the University of Sydney.
Kifaayat
AI Engineer / AI Product Manager. Shipped 7 production agentic automation systems over live marketplace data.
Graduated
Master of Computer Science, Data Science & AI, University of Sydney. Distinction, WAM 76.5.
Founding AI Engineer, Rumblings
Agents, evaluation and the multi-tenant platform behind recommendations for brand and marketing teams.
Experience
Production AI built and owned as both engineer and product manager, on enterprise engineering foundations.
Founding AI Engineer
- Evals Found the weekly customer brief stating false competitor facts while passing every integrity check. Wrote a 17-check evaluation harness before changing anything, then designed the replacement agent workflow with an independent claim checker.
- Trust Built the checks that make model output safe to ship to a client: gates at the source and before delivery, integrity checks before any model call, and provenance on every artefact.
- Platform Designed the multi-tenant Postgres platform from schema to production. Client isolation is tested on every merge, and every recommendation can be reproduced from its run record.
- Product Work directly with the Chief of AI and founders, turning loosely defined requirements into shipped features, including tools the founders run without an engineer.
1 runCustomer onboarding, down from days of manual work
17 checksEvaluation harness written before any fix
RLSClient isolation tested in CI on every merge
TraceableEvery recommendation reproducible from its run record
AI Engineer / AI Product Manager
- Product Assessed workflows across the order and listing lifecycle to identify where automation and agentic AI would deliver measurable benefit, then shipped 7 production automation systems.
- Agents Designed multi-agent LangGraph workflows with agent roles, task routing and conditional escalation over live marketplace data, plus an MCP server that gave models one safe way to query backend data.
- RAG Built RAG systems over marketplace and seller content using pgvector, with confidence gating, citations, and guardrails routing ambiguous cases to human review.
- Roadmap Owned the AI roadmap with business and customer-experience teams, prioritising work by business impact and turning ambiguous requirements into shipped features.
- Quality Established evaluation and observability across factuality, hallucination detection, and task completion; reported outcomes to product and business stakeholders through Power BI over BigQuery.
7Automation systems shipped to production
60%Manual review cut through guardrails and human-in-the-loop routing
43%Listing friction reduced across the marketplace lifecycle
MCPServer standardising how LLMs queried live backend data
Technology Analyst
- DevOps Built and maintained CI/CD pipelines across Azure and AWS, with containerised Docker and Kubernetes workloads deployed across dev, UAT, and production for enterprise clients.
- Delivery Owned a client migration from SVN to GitHub: about 1,200 repositories across roughly 40 teams, delivered on time over four months in phased waves with full history preserved.
- Data Delivered ETL pipelines and REST API integrations with emphasis on reliability, monitoring, and compliance in regulated environments.
- Liaison Acted as technical liaison between engineering teams and clients, resolving production escalations across APIs, authentication, and cross-system data issues.
3 yrsEnterprise cloud and DevOps delivery
1,200Repositories migrated across 40 teams, on time
CI/CDPipelines and containerised delivery on Azure and AWS
2 awardsInfosys Insta Award and RISE Team Award
Selected Work Gallery
Case studies from client work, plus agentic pipelines, retrieval systems and evaluation tooling. Pick a card to open the notes.
Education
Postgraduate computer science study focused on data science and AI.
Master of Computer Science - Data Science & AI
University of Sydney | Graduated December 2025 | Distinction, WAM 76.5
Let’s build AI products that ship.
Open to AI engineering and AI product management roles. Happy to talk about agent systems, evaluation, and turning a messy workflow into an AI product with a number attached.