Celect is rethinking how hiring teams review applications, assess evidence, and identify inconsistencies without reducing candidates to a keyword score.

About the role

We are looking for an AI product engineer to make Celect's application intelligence accurate, explainable, fast, and genuinely useful to recruiters. You will own the application-analysis agent: multimodal resume understanding, structured evaluation, fraud and inconsistency signals, model orchestration, prompts, evals, and the product experience around its results.

This is product engineering work, not a research-only role. You will turn frontier models into dependable systems that hiring teams can understand and review.

What you will do

• Build and evaluate multimodal agent workflows using models available through AWS Bedrock.
• Design structured outputs that turn model reasoning into clear, reviewable hiring evidence.
• Develop representative evaluation sets and measure quality, latency, and cost over time.
• Improve prompt, tool, and schema design based on real failure cases.
• Add guardrails for privacy, reliability, bias risks, and human review.
• Work with product and engineering to ship improvements directly into recruiter workflows.

What we are looking for

• Experience shipping LLM-powered features into real products.
• Strong TypeScript or comparable backend engineering skills.
• A practical understanding of model evaluation beyond a polished demo.
• Comfort debugging the complete path from a document upload to a validated database result.
• Good judgment about latency, cost, privacy, uncertainty, and model limitations.
• Clear communication with both technical and non-technical teammates.

Helpful experience

Experience with AI SDK, Amazon Bedrock, Anthropic models, document understanding, tool use, agent loops, retrieval, trust-and-safety systems, or human-in-the-loop products is especially relevant.