Jyotishmoy Deka
Jyotishmoy Deka is an individual full-stack engineer and independent developer/creator, not an incorporated company. He builds scalable web applications, concurrent backend services, low-level systems software, technical deep-dives, and developer tools, with a focus on React/Next.js, Go, Python, Node.js, and AI tooling.
At a Glance
- Individual software developers and AI coding-agent users
- Engineering teams building LLM applications, agents, RAG pipelines, and internal AI tools
- Teams using GitHub pull requests and automated code review
- Web, backend, and systems-programming learners and practitioners
AI Tools by Jyotishmoy Deka
(2)
AgentTM
Codex Session Recorder and Restore
TraceLLM
Open Source AI Observability Platform
Discussions
No discussions yet
Be the first to start a discussion about Jyotishmoy Deka
Latest News
AgentTM launched on Product Hunt as a local-first time-travel debugging tool for AI coding agents; Jyotishmoy Deka described its timeline, checkpoints, diffs, isolated restoration, Codex chat, and encrypted-backup capabilities.
AgentTM repository updates added a product video, Product Hunt badge, and a deployment-bug fix.
AgentTM v0.3.4 was released on GitHub.
AgentTM v0.3.3 was released on GitHub.
Products & Services
A local debugging tool and desktop/CLI product that records managed Codex sessions and workspace checkpoints, provides timelines, historical file browsing and diffs, and restores earlier files into isolated Git worktrees. The CLI/browser dashboard supports Windows, macOS, and Linux; the native desktop preview is Windows-only.
An observability platform for AI applications covering sessions, traces, spans, events, errors, token usage, latency, metadata, and optional prompt/output content, with a Node.js SDK and OTLP/SigNoz export.
A squad of specialized AI agents that collaboratively audits GitHub pull requests for security vulnerabilities, documentation quality, and code health.
A BitTorrent v1.0 engine built from scratch in Go for download throughput and resilience to network volatility.
Market Position
AgentTM is positioned as a local-first complement to Git rather than a replacement: Git handles commits, branches, and code review, while AgentTM adds managed Codex activity timelines, workspace observations, historical files, checkpoint diffs, and isolated worktree restoration. TraceLLM targets the AI-observability space with product-level traces plus OTLP/SigNoz export. Observer targets automated PR review, differentiating through a fleet of specialized agents for security, documentation, and code health.
Leadership
Founders
Jyotishmoy Deka
Full Stack Engineer at Cogneo Technologies and creator of TraceLLM. Experienced in building scalable systems with React, Go, and Node.js. Previously at Tax Hummer.
Executive Team
Jyotishmoy Deka
Independent full-stack engineer, developer, and creator
B.Tech in Electrical Engineering from Tezpur University; portfolio-listed experience includes Full Stack Engineer at Cogneo Technologies, founding engineer at Tax Hummer, a project internship at IIT Guwahati, freelance web development, and creator of AgentTM and other public projects.
Founding Story
Built by Jyotishmoy Deka to solve the visibility gap in production AI applications, where understanding failure modes and performance bottlenecks is often difficult.
Business Model
Revenue Model
AgentTM is presented as an early public beta with downloadable CLI/browser and Windows desktop builds; public materials identify free options but do not describe a paid monetization model. The portfolio and public repositories primarily represent personal/open-source developer work rather than a disclosed company revenue business.
Pricing Tiers
Product Hunt lists AgentTM under Free Options; no paid tiers were identified in the reviewed sources.
Target Markets
- Individual software developers and AI coding-agent users
- Engineering teams building LLM applications, agents, RAG pipelines, and internal AI tools
- Teams using GitHub pull requests and automated code review
- Web, backend, and systems-programming learners and practitioners
- Developers investigating and safely recovering changes made by AI coding agents
- Teams instrumenting and debugging LLM-powered applications
- Engineering teams seeking automated GitHub pull-request security, documentation, and code-health reviews
- Web and backend application development involving bookings, coding evaluation, CRM workflows, and dashboards
- Learning and demonstrating systems programming, distributed systems, and network protocols