Recruitment technology has a credibility problem: everyone claims AI matching, and every recruiter has been burned by a tool whose "match score" was a keyword count in a trench coat. Building HRTech that practitioners actually adopt means solving the real workflow, which is high-volume, deadline-driven, and trust-scarce, instead of the idealized one in the pitch deck. This is the market I operated in daily, through to an acquisition.
Built and exited: WisOwl AI
I co-founded WisOwl AI, an agentic hiring platform for the Indian job market. I designed and implemented its embedding-based semantic matching engine on FAISS and Supabase pgvector, and shipped autonomous recruiter agents that match candidate supply and role demand in real time. Traction before the acquisition: 10,000+ organic jobseeker signups and 10+ recruitment-firm and startup customers with zero paid marketing. I owned that P&L.
What that means for your HRTech product
- Matching quality is an evaluation problem. I'll help you define what a "good match" measurably is for your users, build the golden datasets, and stop shipping on vibes.
- Two-sided liquidity is sequencing. Candidates without recruiters churn; recruiters without candidates never activate. I've worked through the cold-start problem with real constraints and can pressure-test your plan.
- Recruiter workflow is sacred. Features that demo well but add clicks die in week two. I design around the ATS, the phone, and the WhatsApp thread that recruiting actually runs on.
- Agents need guardrails. Autonomous outreach and screening carry real fairness and brand risk. I build the human-in-the-loop checkpoints before they're needed.
Earlier, I spent eight years at CaaStle running growth product across a ~$50M+ ARR B2B SaaS portfolio, which is useful scar tissue for HRTech companies selling into enterprises. I take fractional roles, advisory retainers, and scoped builds; I'm happy to start with a teardown of your matching or activation funnel.