Nikhil Verma

← All writing

About

Nikhil Verma

CTO, AI systems engineer, and researcher

I’m Nikhil Verma, CTO and co-founder of The Non-Functional Co. in Pune, India. I lead the architecture, AI systems, and product engineering behind Stance, a compliance automation platform for regulated financial institutions.

I’ve spent more than 19 years building and scaling consumer and enterprise software. My work has taken me across India, Sweden, and the United Kingdom, collaborating with colleagues from more than 20 nationalities. I specialise in turning ambiguous technical problems into products with measurable commercial outcomes.

Experience

The Non-Functional Co. — CTO & Co-Founder

2024–present · Pune, India. I lead architecture, AI systems, product engineering, and technical delivery for an enterprise AI-compliance platform used in regulated finance.

Ollion — Principal Developer

2021–2024 · Pune, India. I led frontend engineering for CodePipes, a multi-cloud DevOps platform, and Stance, which later became The Non-Functional Co. I led a cross-functional team of ten engineers and drove the move from Vue 2 to Vue 3 and TypeScript while continuing feature delivery.

Bumble Inc. / Badoo — Engineering Lead

2013–2021 · London, United Kingdom. Over eight years I progressed from engineer to Mobile Web Team Lead and then Frontend Core Team Lead. I led teams of up to ten engineers and helped unify frontend architecture, build systems, design systems, experimentation, and quality practices across web products serving tens of millions of users.

Opera Software and AOL

2008–2013 · Gothenburg and Bangalore. I worked on consumer web products at Opera Software and AOL after earning a B.Tech in Computer Science.

Research

Active Context Compression: Autonomous Memory Management in LLM Agents

My sole-author research explores an agent that autonomously compresses its own context, inspired by slime-mold exploration. It achieved a 22% token reduction on SWE-bench Lite without reducing accuracy.

What I work on

  • Reliable and deterministic LLM-agent systems
  • AI compliance for regulated financial institutions
  • Context compression and long-running agent memory
  • Engineering leadership, architecture, and team building
  • TypeScript, Vue, distributed systems, Kubernetes, and Temporal

Elsewhere