A Rust-native terminal coding agent built around runtime-owned policy decisions, recoverable workflows, and evidence-based verification.

Rust Ratatui SQLite · Documentation · Releases

An Arch-based Linux ecosystem focused on desktop experience, system tooling, and the performance/elegance balance.

Linux Arch i3 · GitLab projects

Self-hosted organizational webmail that connects to existing IMAP/SMTP infrastructure. It deliberately avoids third-party AI analysis of message content.

PHP Laravel Livewire · Setup guide · Security policy

• Octate — code-review tooling combining local static analysis with optional model-backed review. TypeScript Node.js
• SLMGen — tools for dataset analysis and small-language-model fine-tuning workflows. Python FastAPI TypeScript
• Runstead — a hosting-business platform concept with an interactive browser-first simulation; infrastructure actions and payments are simulated. Next.js TypeScript

Project capabilities and release status vary. Follow the repository README, issue tracker, release notes, and CI workflow for current implementation status.

I prefer implementation evidence over vanity metrics. These links go directly to each project's automation and release history; they do not imply that every workflow is currently passing.

• Architecture before implementation. Keep domain boundaries understandable and explicit.
• Measure before optimizing. Prefer reproducible evidence over performance claims.
• Fail visibly. Return actionable errors instead of hiding broken or unsupported paths.
• Treat inputs as untrusted. Validate external data, model output, repository content, and user-supplied configuration.
• Make recovery part of the design. Cancellation, checkpoints, rollback, and diagnostics are not afterthoughts.
• Test behavior, not just code paths. Use unit tests for local invariants, integration tests for boundaries, and end-to-end tests for critical user journeys.

Systems: Rust, Go, Linux, shell tooling
Application engineering: TypeScript, React, Next.js, Flutter, PHP
Infrastructure: Git, CI/CD, deployment automation, databases
Applied research: small language models, dataset pipelines, evaluation workflows

I value focused issues, reproducible bug reports, reviewable pull requests, and documentation that reflects what the code actually does.

• Browse my repositories
• Report a bug or request a feature · use the relevant project's issue tracker for project-specific reports
• Explore Tonmoy Infrastructure

_{Maintained by Eshan Roy. Prefer verifiable behavior and current project documentation over unverified feature or performance claims.}