AI Product Manager

Responsibilities:

• Experimenting with different models, prompts, context, and system configurations to determine what produces the best user experience.
• Creating evals that measure whether an AI system is genuinely improvingnot merely producing outputs that look good in a few hand-picked examples.
• Running hundreds of A/B tests across prompts, models, product flows, and AI behaviours.
• Defining and tracking the user and business metrics associated with these systems, including quality, engagement, conversion, retention, latency, and cost.
• Studying real customer interactions to identify failure modes, unmet needs, and opportunities to improve the product.
• Building systematic feedback loops between user behaviour, evaluation results, and product changes.
• Working with engineering and data teams on model selection, fine-tuning, datasets, guardrails, and production performance.
• Balancing output quality with speed, reliability, cost, and the constraints of serving millions of users.
• Developing a point of view about when better prompting is sufficient, when the product experience needs to change, and when fine-tuning or deeper model work is justified.

Requirements:

• 1+ years in Product, Growth, Analytics, Growth Marketing, or an early-stage company, with evidence that you have owned and shipped measurable outcomes.
• Strong product and user judgement. AI products rarely have one objectively correct outputyou can recognise what feels helpful, natural, accurate, and trustworthy for the user.
• Analytical and engineering thinking. You can break AI quality into measurable dimensions, design evals, investigate failure modes, compare model behaviour, and optimise systems methodically. PMs here need to be technically fluent.
• Independence with data. You know which questions to ask, can retrieve and analyse the data yourself, and don't depend on a data analyst for every answer.
• First-principles problem-solving. You don't blindly copy existing playbooks; you break ambiguous problems down, find the root cause, and develop your own point of view.
• A builder's mindset. You take a problem from uncertainty to hypothesis to shipped product and remain accountable for what happens afterwards.
• Comfort with speed and experimentation. You would rather run a credible test and learn than spend weeks trying to perfect a plan.
• Technical fluency. You don't need to be a software engineer, but you should be comfortable working closely with models, prompts, evals, datasets, and engineering systems.
• Clear communication and high ownership. You can bring people along, disagree directly, and keep work moving without continuous supervision.

Good to Have:

• Very fluent in using AI tools in your own workflow to move faster - Very important.
• Hands-on experience shipping AI / LLM-powered features.
• Strong engineering pedigree IIT/IIM/top-tier institute background is a plus, not mandatory.
• Experience at a consumer app, consumer AI, or high-growth product company.