RansahAI — AI-Powered Hiring Automation Platform
Eval-Driven AI Hiring at 0→1
The Challenge
Hiring teams were drowning in manual screening work with no reliable way to use AI in the loop — existing tools were either too black-box to trust or too rigid to configure. There was no eval layer, no governance, and latency on AI-backed workflows made real-time use impractical.
My Approach
Designed and shipped multi-agent screening workflows (LangGraph/n8n) with golden-dataset eval pipelines (DeepEval), prompt optimisation, retry logic, and semantic caching — then layered credit-based pricing and PLG onboarding (checklist, demo data seeding, role-aware nudges) on top to drive activation.
Impact
Pilot latency dropped 40%; activation and time-to-first-value improved measurably; credit-based pricing model built a clear path to monetisation with unit economics tracked at the token level.
Key Highlights
- Full-stack ownership — AI architecture, pricing model, growth loops, and governance in one role
- Golden Dataset eval baselines via DeepEval before any workflow ships
- GraphRAG + Neo4j + Redis retrieval stack — not just API calls
- GCP→AWS migration driven by GDPR compliance
- DSPy for programmatic prompt optimisation
- PQL→SQL pipeline with explicit handoff criteria to Sales