Technical Product Manager, AI Submission Ingestion Platform – India

About the role

What you'll do

• Position Summary:
• As Technical Product Manager you own a product area of the platform end to end: the problem, the outcome, the spec, the quality bar and the release. You know the architecture, data and AI well enough to make real trade-offs with engineers, and you are close enough to underwriting and operations to know which trade-offs matter. Within the platform strategy and portfolio priorities set by platform leadership, you decide what gets built in your area, in what order, and how we will know it worked. You work as a peer with TPMs in the US and India, each owning a different area of the same platform.
• In Hyderabad, this role's center of gravity is engineering proximity. You own a product area of the platform, and you sit with the squads that build it. You make sure what ships in your area is specified precisely, evaluated honestly, released cleanly for users in every region, and moves the outcomes it was built for.
• Product area ownership: Own the roadmap, backlog and outcome measures for your area within the platform strategy. Make the scope and priority calls, keep the backlog ready two sprints ahead, and answer engineers' questions the same day.
• Business partnership: Work directly with underwriters, operations users and business SMEs on field definitions, rules, accuracy targets and ground truth for your area. Bring regional needs into your roadmap firsthand.
• Specification: Write stories engineers build from without coming back with questions: acceptance criteria as Given/When/Then scenarios, API contracts, event flows, data models, and explicit non-functional requirements.
• Evaluation and ground truth: Build and maintain the golden datasets and evaluation runs for your area. Analyse misses field by field, decide whether the fix is configuration, rules, prompt or model, and turn reviewer corrections into training and evaluation signal.
• Simplest workable path: Rules first, then deterministic tools, then a single model call, and an agent only when the step below fails. Hold the line on cost, latency, explainability and human escalation points.
• Release readiness: Own the readiness gate: acceptance scenarios met, evaluation thresholds passed, regression scope covered by business risk, performance, resilience and access control verified. Validate key functionality in UAT and production yourself. Track predictability, flow and escaped defects, and run retrospectives that change something.
• Global users: Coordinate UAT and rollout with users in North America, EMEA, APAC and LATAM. Publish release notes in business value, not ticket numbers.
• Build with AI: Prototype workflows and screens, analyse evaluation output, and read code and logs to answer your own questions. Show a working prototype, not just a slide.
• Technical Skills:

Nice to have

• Commercial P&C insurance: submission intake, clearance, underwriting workbenches, rating or policy administration.
• Intelligent document processing or extraction platforms, ontologies or knowledge graphs, canonical data models.
• Cloud platforms (Azure preferred), event streaming, workflow orchestration.
• Prior hands-on experience as an engineer, data scientist or business analyst.
• Agile delivery at scale in a regulated enterprise, across time zones.