I'm Eshan Roy, Founder & CEO of Tonmoy Infrastructure & Vision. I turn ambitious engineering ideas into dependable software, working across autonomous AI, Linux ecosystems, developer tooling, and infrastructure.
|
01 / THINK IN SYSTEMS I start with first principles: clear architecture, explicit boundaries, and an understanding of how components behave together. |
02 / BUILD THE FOUNDATIONS I work close to the runtime, tools, platforms, and infrastructure that make ambitious products practical. |
03 / SHIP WITH INTENT I value secure defaults, predictable behavior, useful diagnostics, and experiences that respect the people using them. |
|
M31A — Autonomous software engineering An open-source engineering agent focused on repository understanding, planning, governed tool execution, and verifiable development workflows. Repository · Product · Docs |
An independent Arch-based operating-system ecosystem spanning desktop experiences, system applications, packaging, and developer-first workflows. |
|
Exploring agent-driven desktop workflows, browser interaction, automation, and user-controlled execution. |
Octate — Terminal-native code review Developer tooling that brings repository analysis and code-review workflows closer to the command line. |
A collection of experiments and products exploring the edges of software, infrastructure, and developer experience.
|
Privacy-conscious webmail designed to work with existing mail infrastructure. A self-hosting-oriented approach to organizational email. MAIL · PRIVACY · IMAP / SMTP |
Tooling for small-language-model experimentation and dataset workflows, from preparing source data to building useful model pipelines. AI · DATASETS · MODEL TOOLING |
|
A hosting-business management platform exploring the workflows behind customers, services, billing, and infrastructure operations. HOSTING · OPERATIONS · PLATFORM |
A lightweight browser exploration focused on a clean, approachable browsing experience without unnecessary complexity. BROWSER · DESKTOP · USER EXPERIENCE |
|
Exploring the Model Context Protocol ecosystem and assistant tooling: connecting AI applications with external capabilities through structured tool interfaces. MCP · AI INTEGRATIONS · TOOL ECOSYSTEM |
|
Each project is evolving independently. Visit its repository for implementation details, current capabilities, and contribution opportunities.
|
|
|
COMMIT PULSE · CONTRIBUTION HISTORY
Charts are generated from public GitHub activity. External chart services may cache results before refreshing.
Architecture with intent. Clear boundaries, explicit contracts, and systems that can evolve.
Security by design. Least privilege, explicit trust boundaries, and safe defaults.
Evidence over hype. Reproducible tests, observable behavior, and honest project status.
Reliability is a feature. Cancellation, recovery paths, diagnostics, and predictable failure modes.
Open by default. Useful documentation, reviewable changes, and constructive collaboration.



