Sai Mageshvar.

Engineer / Team lead / Tool builder

Leading teams.
Building what
comes next.

Engineering judgment,
amplified by AI.

Explore selected work
✳ Chennai, India / A working notebook by Sai
find the friction.
build a better way.Pencil drawing of a little workshop sorting a large stack of test sheets into a selected few
FIG. 01Small tools. Meaningful differences.
01

Selected work

Built from real problems

A few things I’ve built to make
engineering work a little better.

from the notebook
01 / Test impact analysisOpen source

codeflux

Run the tests your
change affects.

20→7min

CI runtime at Chronus

a small changetrace its impactjust these tests
6 tests in the sample suite

An illustrative sketch of test selection.

Built an open-source Rust CLI that uses runtime tracing to identify tests affected by changes in Ruby projects. Adopted by Chronus to run affected tests and reduce CI runtime.

Rust / Ruby / Open sourceView source
02 / Code reviewOpen source

diff-easy

Less noise.
A clearer picture.

1234a little more clarity

Illustrative file order.

500+downloads in
15 days

Extended an open-source diff viewer into an npm tool for reviewing AI-agent work across multiple repositories from one tabbed hub. It replays the exact diff from every agent turn and uses AI to group changed files by intent. The package reached 500+ downloads in its first 15 days.

npm / AI toolsView source
03

herdr-auto-pilot

Agent Automation / Python / Claude Code

Built agent automation that keeps Claude Code sessions moving through routine approvals and completion checks, with decision logs and escalation when human input is needed.

04

rich-prompt

Open Source / TypeScript / Claude Code Mods

Built a Claude Code mod adding word autocomplete and familiar editor behaviors to the prompt box, adopting the mod system early to make everyday agent interactions smoother and more effective.

Explore the project
05

Annotate for Agent

VS Code / AI Tools

Published a VS Code extension on the Marketplace that turns inline code-review annotations into structured feedback for AI agents.

More from the workbench (10 projects)

Bunch of Utils

VS Code extension (v6.6, published on the Marketplace) with 10+ productivity tools—smart test execution, fuzzy file mentions, JSX prop transformation, voice notes, Git modified-file search, an in-editor notes manager, and text-manipulation utilities.

VS Code / Extension / Productivity

RAG-Valluvar

RAG-powered web app providing life guidance from Thirukkural—matches user scenarios to classical wisdom using LLM + vector search.

AI / RAG / LLM

TaskFoldr

Folder-based todo list built for agents—each task is a self-contained folder holding screenshots, references, and logs.

Productivity / Agents

task-manager-mcp

MCP server built with FastMCP that gives AI agents structured task management, created during the early days of MCP.

MCP / FastMCP / AI

copilot-survey-tool

Internal GitHub pull-request survey tool measuring Copilot effectiveness across the organization.

GitHub / Copilot / Analytics

stat-story

Automated agent generating Remotion infographic videos from global statistics for YouTube content creation.

Remotion / Video / Automation

pretty-git

CLI wrapper for Git with a tree-based branch hierarchy and simplified log, stash, and diff commands.

CLI / Git / Developer Tools

Naem.in

Baby-name discovery website with shortlisting, polling, and AI-powered name fusion.

Web / AI / Social

NeverEmpty

Android app that tracks usage patterns of household items and sends proactive refill notifications.

Android / Mobile

chapters-of-us

Interactive story generation with rooms, scenes, and objects, built with Phaser.

Phaser / Game / Interactive
02

The way I work

People, systems & the craft
Hands on, eyes on the bigger picture

A team to lead.
Releases to ship.
Still building.

Lead a six-person team, including two senior engineers, a senior QA engineer, and a tech lead.

Built Lighthouse to coordinate team priorities and multiple releases while continuing to contribute hands-on, bringing delivery progress and follow-ups into one view.

The tools I build often start here: something that could be clearer, a repeated task, a little friction worth removing.

keep the whole
picture in viewthe teamthe craftthe releases
FIG. 02 / LighthouseA little perspective for the everyday.
01.

Find the real problem.

Start with the work and the people doing it.

02.

Build a useful answer.

Choose the tool that helps the work move forward.

03.

Keep your judgment.

Use AI to go further. Stay responsible for what ships.

03

The longer story

Experience / Chronus

Engineering depth.
A wider perspective.

AI-Accelerated Engineering Leader with 9+ years of hands-on experience.

Cut release cycles from six weeks to one and built an in-house software factory that completes Jira work from written specs on cloud runners.

Now manage a six-person team and lead the migration of hundreds of pages from a Rails monolith to an API-first SPA. Built systems serving 200+ enterprise organizations and shipped Chronus's first Generative AI product.

July 2024 — Present01

Team Lead (Hands-on)

Chronus

  • Championed AI-driven development organization-wide—rolled out Claude Code and Copilot across engineering teams, compressing release cycles from 6 weeks to 1 week.
  • Added autonomous delivery capacity alongside engineering teams by building a software factory that launches parallel cloud workers and takes Jira-backed specs through implementation, testing, and risk-controlled merge.
  • Took the CTO's concept from design to company-wide rollout, extending OpenAI's open-symphony with Jira integration and reusable AI SDLC skills while owning orchestration from intake to merge.
More about this role
  • Led the company's first Generative AI product release, designing and optimizing production-grade prompts and shaping the enterprise AI roadmap.
  • Reimagined a complex user-facing feature end to end in 8 hours using agentic toolchains—work originally scoped at 6–7 weeks—delivering a radically improved UX through rapid prototyping.
  • Reduced CI runtime by 65% (20 to 7 minutes) through company adoption of codeflux to run only tests affected by a change, saving engineers hours of wait time daily.
  • Lead a major platform migration from the monolith to an API-first backend and single-page application. The release spans hundreds of pages, requires backend changes for the new architecture, and includes new skills that identify common page archetypes and generate pages faster.
  • Lead a six-person team, including two senior engineers, a senior QA engineer, and a tech lead.
  • Built Lighthouse to coordinate team priorities and multiple releases while continuing to contribute hands-on, bringing delivery progress and follow-ups into one view.
  • Led yearly product roadmap estimation and planning for the team, aligning engineering capacity with business priorities and stakeholder expectations.
  • Served as an active member of the Culture Club, the organization-wide body that drives employee engagement, eNPS, and people-focused initiatives.
July 2021 — June 202402

Senior Software Engineer I & II

Chronus

  • Directed one of the largest releases in company history—a 9-month, 2,000+ file platform overhaul coordinating 7+ engineers across teams, delivered on schedule.
  • Delivered multiple major releases spanning multiple quarters with zero critical defects.
More about this role
  • Designed microservice architectures with REST APIs serving 200+ organizations, including Fortune 500 and MAANG companies.
  • Architected a plug-and-play admin analytics dashboard in React—surfacing organizational KPIs, health metrics, and strategic direction—adopted by all enterprise clients.
  • Conducted technical interviews for experienced hires and campus recruits, evaluating candidates across experience levels and contributing to hiring decisions.
June 2017 — June 202103

Software Developer I & II

Chronus

  • Developed and maintained core product modules with Ruby on Rails and MySQL, serving thousands of enterprise users.
  • Improved system reliability through targeted refactoring and data-consistency enhancements, reducing production incidents.

The toolkit

Tools change.
Curiosity stays.

AI & Agentic

Claude Code, GitHub Copilot, nanobot, NanoClaw, RAG, Prompt Engineering, AI-Augmented SDLC, Agentic Workflows

Backend

Ruby on Rails, Python, REST API Design, Microservices, MySQL, Redis

Frontend

React, JavaScript, jQuery, HTML/CSS

DevOps & Tooling

Git Worktrees, CI/CD, Agent-Browser Testing, Docker

Leadership

Team Management (6-person team), Roadmap Planning, Performance Management, Technical Hiring, Cross-Functional Collaboration

A few milestones

Learner of the Year

Chronus—consistently ranked #1 on the company's learning leaderboard for over a year, embodying the core value of Continuously Learn; recognized by the CEO with a dedicated award. View on LinkedIn.

Chronus Hackathon & Conference Winner

Chronus, 2018—part of a three-member team that integrated the product with Google Assistant for voice-based workflows. Delivered a winning talk on database sharding at the first-ever Chronus conference.

Best Individual Performer & Technical Excellence

Chronus—recognized multiple times at company town halls for individual excellence, technical leadership, and team achievements.

Education

B.E. Computer Science & Engineering

College of Engineering, Guindy · 2013 — 2017

Higher Secondary

DAV Gopalapuram, Chennai · 2011 — 2013

The next conversation

Good things start
with a hello.

saimageshvar@gmail.com
a note is always welcome

Always curious about agentic workflows, AI-native products, and systems that ship fast.