Satya Borgohain

Building Kensa to improve agent reliability. I write about evals, engineering, and closing the gap between demo and production. If you find something interesting, say hello or grab some time.

Writings

Aug 17, 2026

Does your intent survive?

1 min

Jun 1, 2026

Devloop: closing the loop

4 min

May 22, 2026

Happy paths are edge cases now

3 min

Apr 8, 2026

Kensa: an open source agent eval harness

2 min

Apr 2, 2026

Why build the agent verification layer

2 min

Mar 15, 2026

Physician disagreement in healthcare evals

5 min

Mar 9, 2026

The half-life of benchmarks

5 min

Feb 16, 2026

Blurt: talk to your agents

2 min

Feb 12, 2026

Human review is the bottleneck

5 min

Jan 14, 2026

10x coding agents

7 min

Sep 12, 2025

Directive for Claude

2 min

Jun 30, 2025

Multimodal RAG

8 min

Jun 23, 2025

The paradox of infinite context

4 min