# Madhusuthanan B (thecompiledmind)

> Senior Staff Engineer at Falkor (formerly Kongsberg Digital), working on AI platforms and agents. Hands-on architect with 13 years in system design, core engineering and leadership. Crafts the layers below the product, for humans and agents.

I design the systems underneath the product that people and AI agents both depend on. I was part of the core team that built the Reasoning Engine, Falkor's in-house agentic framework and harness, and I now own it end to end. It powers the Asset Copilot in the Falkor platform, other teams across Falkor use it to build their own custom agents, and AI agents have run on it in production for over two years. I work AI-native: specs before code, AI for speed, people for judgement.

## Pages

- [Home](https://thecompiledmind.dev/): overview, career story, toolbox, manifesto and contact
- [The Substrate Manifesto](https://thecompiledmind.dev/manifesto): a manifesto for AI-native engineering, in seven lines

## Work, newest first

- [The AI layer for the digital twin](https://thecompiledmind.dev/#ch-engine): Falkor, now. The Reasoning Engine, an agentic framework and harness for industrial digital twins (oil and gas). Agents and multi-agent orchestration around one agentic loop, working with any model provider.
- [Condition monitoring for Kongsberg Maritime](https://thecompiledmind.dev/#ch-maritime): Kongsberg Digital, from 2019. Part of the team that built Vessel Insight (fleet data to the cloud). Tech lead for the vibration-analysis applications, on-premise and in the cloud, with a team across Norway and India; introduced behaviour-driven development.
- [The manufacturers portal for ItemMaster](https://thecompiledmind.dev/#ch-itemmaster): 2018 to 2019. First employee in Bangalore; built the web app where brands managed their product content.
- [Client-facing products for Charles Schwab](https://thecompiledmind.dev/#ch-schwab): 2013 to 2018, with Wipro. Five client products, built as micro frontends from 2016; test-driven development.

## The Substrate Manifesto

1. Design for people and agents. Every system now has two kinds of users.
2. Behaviour first. Start from the behaviour you expect, then write the code.
3. Build the layer, not the feature. A good layer keeps paying off for every team that builds on it.
4. AI for speed, people for judgement. AI moves fast on the rest; architecture, quality and correctness stay with people.
5. Craft over output. When AI makes code cheap, knowing what good looks like, and caring enough to get there, is what's left to compete on.
6. Ship small, ship often. Small, frequent releases are easier to trust and easier to fix.
7. Own it in production. Building it is the start; running it is the job.

## Toolbox

In each group, tools before "also" are in daily use.

- AI and agents: Claude Code, GitHub Copilot, MCP, Skills, OKF, Agent memory, LLMs, OpenAI APIs, Responses API, RAG, Qdrant; also Cursor
- Platform and backend: TypeScript, Node.js; also .NET Core, REST APIs, MQTT, Node-RED
- Data and cloud: Redis, MongoDB, Docker, Kubernetes, OpenTelemetry, Azure; also On-premise deployments
- Frontend and data viz: Angular; also JavaScript, D3.js
- How I work: AI-native engineering, LLM Wiki; also C4 modelling, BDD, TDD
- Exploring (not in production yet): WebMCP, System 1 models

## Contact

- [LinkedIn](https://www.linkedin.com/in/madhusuthanan-b/)
- [GitHub](https://github.com/Madhusuthanan-B)
- [X](https://x.com/thecompiledmind)
- [Medium](https://madhub081011.medium.com/)

## Optional

- [Mindbase](https://github.com/Madhusuthanan-B/Mindbase): open-source personal knowledge base on the OKF standard, with agent skills for Claude Code