I build AI assisted products for customer experience and contact centres. Ambiguous requirements come in, shipped automation goes out. I decide how the model behaves inside the product, where it hands back to a human, and what we measure to know it is working.
Expertise
I sit between technical execution and customer experience, turning discovery into PRDs, designing how AI shows up inside the product, and prioritising the roadmap so the right things ship first. Today I own product for a global lighting client, Signify, at Tech Mahindra. Before that I drove enterprise CXM and CCaaS for Deutsche Telekom, Walmart, Emma Sleep, Wayfair and LG at Sprinklr.
My happy place is the messy middle. Gap analysis, impact assessment, routing logic, agent facing UX, and the prompt design that makes an in platform model behave the same way twice. I am pursuing an Executive MBA at IIM Indore alongside the day job, and I build things on weekends because reading about a model and shipping one turn out to be different skills.
Companies and clients: Tech Mahindra, Signify, Deutsche Telekom, Walmart, LG, Emma Sleep, Wayfair, Sprinklr, SpOvum, Alten. Tools and schools: Anthropic API, Claude, GPT-4, Sprinklr CCaaS, Next.js, Supabase, IIM Indore, IIT Goa.
Selected work
The first four are mine end to end, from problem to production. The last two are enterprise programmes where I owned product definition and the rollout.
A churn and expansion forecasting system for B2B SaaS. Structured extraction pulls signals out of unstructured account activity, scores account health, and flags the accounts a CSM should call this week. Built with a synthetic data pipeline so the whole thing stays inside GDPR and DPDP constraints. This is my Executive MBA thesis and I am writing the code myself.
A career advisory product for affluent Indian families, built on one falsifiable hypothesis with a kill criterion written before any work started. Covers jobs to be done segmentation, services unit economics, RICE prioritisation, and a deliberate line between where the model assists and where a human signs off.
Read the caseA privacy first utility suite, currently being repositioned around contact centre and voice tooling. Every tool runs in the browser, nothing a user types leaves their machine. I set the constraint, designed it, built it and shipped it on my own.
Visit the siteA training planner that turns your available days, equipment and goal into a split you can actually follow. Shipped as a working planner first. The model layer that adapts the plan from logged sessions is going in next.
Try itOver 10,000 agents on one deployment. I owned routing logic, automation rules and the escalation paths across both automated and agent handled contacts, then ran A/B tests and adoption dashboards to guide releases. It delivered a 20 percent lift in product utilisation and more than 10 million euro in annual client savings.
Business day aware out of office automation across email, voice and live chat, with prompts engineered so downstream systems can parse the output every time. Plus PII masking rules, a call wrap interface that captures the fields without burying agents, and a vendor evaluation that killed a licence we did not need.
The part most people skip
Most product people who say contact centre mean chat and email. I spent four years at Sprinklr as the voice subject matter expert, covering both inbound and outbound, and the outbound side is where the hard product problems live.
Dialer pacing, campaign design, list strategy, compliance windows, abandon rates. If you are putting an AI agent anywhere near a phone line, these are the constraints that decide whether it works or gets switched off in week three.
I combine product discipline with AI assisted execution, moving from research to release without losing the detail along the way.
Where I have worked
Own product definition and the roadmap for Signify's Sprinklr based CX platform, turning ambiguous business requirements into shipped automation across email, voice and live chat.
Configured AI agent behaviour in production, including PII masking rules and structured response templates.
Specified guardrails and fallback paths, including business day aware out of office logic, so automation degrades safely instead of answering incorrectly.
Engineered prompts for in platform GPT-4 to return deterministic, parseable output for downstream automation.
Ran an 18 day enablement programme for developers and QA on the platform.
Led enterprise CXM and CCaaS deployments for Deutsche Telekom, Walmart, Emma Sleep, Wayfair and LG, running discovery through to user stories, PRDs and production release.
Drove Europe's largest Sprinklr CCaaS rollout at over 10,000 agents, owning routing logic, automation rules and escalation paths across automated and agent handled contacts.
Delivered more than 10 million euro in annual client savings and a 20 percent lift in product utilisation on that deployment.
Acted as voice subject matter expert for inbound and outbound, with depth in outbound dialing and campaign management.
Partnered with product and engineering on gap analysis, API and integration requirements and roadmap prioritisation.
Ran A/B tests and adoption dashboards to guide releases.
Ran the end to end lifecycle of biotech devices, from prototyping and feasibility through iteration on lab feedback to release.
Managed SAP systems and service documentation for in service aircraft, holding data integrity across maintenance records.
Education and recognition
IIM Indore
2025 to 2027, weekend programme for working professionals. Class Representative through Term III.
IIT Goa
2017 to 2021. Team Captain, IIT Goa Motorsports. Second runner up, Stand Up Comedy, BITS Goa.
IBM on Coursera
Product strategy, discovery and delivery fundamentals.
Awarded twice at Sprinklr, in Q3 2026 and Q2 2024, for enterprise rollouts that moved the numbers rather than just closing the ticket.