katyal.ai

Work

How we engage.

Three shapes. Each starts with a conversation about what you'd consider a good outcome — and what would make it not worth doing at all.

Discovery sprint

Two-week scoped engagement to figure out whether a specific AI idea is worth building.

  • Honest assessment of what current models can and can't do for the problem
  • Working prototype or proof point on real data
  • Evaluation harness and a short write-up: build / don't build / build differently

Production build

Multi-week engagements to take a validated idea into a maintainable system.

  • Architecture grounded in current best practices for the model family in use
  • Tests, evals, telemetry, and a runbook your team can operate without us
  • Documented decisions in the same git history as the code, not in a separate slide deck

Direction & advisory

Recurring time with a small team or founder to keep AI work pointed at the right problem.

  • Regular working sessions on scope, evaluation strategy, and tradeoffs
  • Written notes in a shared portal — not just verbal advice that evaporates
  • Honest pushback when something isn't worth doing yet

Not sure which fits? Tell us what you're trying to do — we'll suggest the shape that makes sense, or tell you if nothing does.