AI Product Engineer
HybridKarachi, Sindh, Pakistan
Job Summary
Own the end-to-end product lifecycle for AI features across the GuardPass ecosystem, from discovery through to production shipping. Build working prototypes using React/TypeScript front ends, Python/Node back ends, and LLM APIs rather than relying on documentation. Design and implement LLM-powered systems including conversational interfaces, RAG pipelines, and multi-agent workflows with tool-use integrations. Develop evaluation frameworks to define success metrics and drive iteration, while writing and refining production prompts as versioned engineering artifacts. Collaborate with lead engineers and product managers across UK and Karachi teams to deploy prototypes, instrument features with funnels and latency tracking, and make pragmatic build-vs-buy model selection calls.
Required Qualifications
- 3+ years combined experience across product management and software engineering (in either order — but real experience in both)
- Shipped at least one LLM-powered feature or product to real users — not just demos
- Hands-on coding ability: comfortable in Python and/or TypeScript, working with REST APIs, and shipping full-stack prototypes independently
- Practical experience with LLM APIs (Anthropic, OpenAI, or similar), including structured outputs, tool use / function calling, and context management
- Strong product judgement: you can say no, prioritise ruthlessly, and articulate why a feature matters in commercial terms
- Clear written communication — specs, briefs, and decision docs that others can act on
Desired Qualifications
- Experience with agentic frameworks or patterns (Claude Agent SDK, LangGraph, MCP, or equivalent)
- Experience building and running LLM evals (promptfoo, custom harnesses, or similar)
- Familiarity with modern deployment stacks (Supabase, Railway, Cloudflare Workers, AWS)
- Experience in marketplaces, booking flows, or regulated industries
- Exposure to conversational UX design — voice or chat
- Prior startup or founder experience
- Data analysis skills (SQL, basic statistics) for feature measurement
- Design sensibility — you can make a prototype look credible without a designer
Hiring someone like this?
Get your role in front of qualified candidates on Sorce.