Rishikesh S Software engineer

Rishikesh S

Software engineer in Chennai

Founding engineer at Mando · building software since 2022 · IST, UTC+5:30 · IIT Madras

I build software end to end, from the schema to the interface, and I keep building after hours: developer tools, Chrome extensions, Mac menu-bar apps and a few things for cricket. At work I’m a founding engineer at Mando, lately on concurrent editing and AI pipeline reliability. Before that I spent two years on trading and risk systems at Quantitative Brokers.

rishikeshpavithram@gmail.com · GitHub · LinkedIn

Now
Mando, remote, since Sep 2024
Before
Quantitative Brokers, Jul 2022 – Sep 2024
Studied
Chemical Engineering, IIT Madras

Projects

Things I’ve built, the ones I’d show first at the top.

Full write-ups, with builds and downloads

Back to contents

Work

What I built at work: two roles in full, three earlier ones in a line or two.

Sep 2024 – now 2 yrs 2 mos Remote Stack React, Next.js, TypeScript, Python, FastAPI, PostgreSQL, WebSockets, Chrome Extensions (MV3), FFmpeg, Terraform, GitHub Actions, Google Cloud Run, Cloudflare R2, Stripe Connect

Founding Software Engineer at Mando

I joined as a founding engineer and work across the React frontend, the Python services and the cloud.

Workflow capture. I built active and passive workflow capture. It pairs screen recordings with DOM events, so a capture keeps the application actions as well as the video. A first guide appears in seconds while full generation runs in the background, and the same capture feeds search answers with citations and no-code testing.

scope Search and testing are team efforts. My part is the capture that feeds them. Retrieval, ranking and the test replay engine are other people’s work.

Video editor. I rebuilt the React video editor around one shared timeline for trimming, annotations, redaction and voiceovers. QA found that two tabs editing the same guide silently overwrote each other, so I added optimistic concurrency control and three-way merging. A partner’s 4K video with roughly 8,000 annotations hung the editor; indexing the annotations made it responsive again, and automated performance benchmarks now cover load, interaction, playback and save.

mechanism Every save carries a revision, and under an advisory lock the server refuses a save made against a stale one. In the editor each WebSocket snapshot is merged three ways against the last merged base; a conflict offers Keep mine or Use theirs, and a refused save opens a dialog.

Pipeline reliability. I hardened the AI generation and editing pipelines: explicit state transitions, retries, output validation and fallback models. I fixed races in how edits propagated, replaced polling with WebSocket job updates, and grouped terminal failures by root cause so the team could see why a workflow failed or stalled.

Meeting bot and search. I built a cross-platform meeting bot end to end. It turns a recording into a summary, action items and a record you can search and share. I also extended search with streaming answers, source citations, screenshot attachments and links to the exact workflow step, both in the app and in the browser side panel.

also I moved the Chrome extension from a popup to the side panel, so search sits beside the page you’re working in.

Partners and managed services. I shipped branded apps and Chrome extensions for 3 partners. Partner deployments now run through one Terraform and GitHub Actions setup. I also built managed-services ticketing: requests route to the right specialist, SLA timers count business hours, and expert invoices go out automatically through Stripe Connect.

scope I don’t name customers or partners.
mechanism Monthly invoicing sweeps unpaid hours, builds an Excel workbook in SharePoint via Microsoft Graph and drafts a Stripe Connect invoice, with advisory locks and idempotent retries.

Data integrity. I wrote validation and repair tooling for stored workflows. It checks for invalid timestamps, missing sources and gaps in step order. Running it on production showed legacy asset links still pointing at DigitalOcean Spaces, so I migrated 12,108 records across 681 workflows to Cloudflare R2, keeping stored workflows connected to their media.

Cloud migrations. When our Azure credits ran out I moved us off Azure and retired 12 databases across 4 environments. Later, DigitalOcean’s uptime problems led me to move the backend to Google Cloud Run: both platforms ran in parallel until the DNS cutover, and I documented the deployment and migration procedure. 

mechanism Leaving Azure took two moves: Postgres to Neon, three databases in each of four environments, then the containers to DigitalOcean. All three migrations ran through Terraform and GitHub Actions.
Other work at Mando the copilot, the marketplace, shared libraries, APIs and data safety

My first year went mostly into the copilot (it later became the product’s search) and the expert marketplace. For the copilot I built the answer interface from scratch: answers stream through the Vercel AI SDK and LangChain, inline citations are numbered and de-duplicated, and inference endpoints that had been public now need an API key.

For the marketplace I built a booking flow, expert chat and video calls on Azure Communication Services (the calls later moved to Twilio Video with recording), and Stripe Connect payouts to experts. Alongside it I reworked the Stripe subscription service for seat-based pricing behind customer plans, including seat changes and onboarding.

I extracted the shared UI and design-token libraries, and added the first frontend CI, Storybook tests and per-PR Storybook previews.

I built the expert-training and labelling pipeline: expert answers pass through a verification state machine, and an aggregator merges them into the labels that improve the copilot.

I ran an API quality programme: a CI coverage report found that lazy routers exposed only about 15 of roughly 925 routes to tooling, and I moved the generated client into a shared typed package that the frontend and the extension both use.

After a backfill brought deleted docs back in production, I standardised soft deletes behind one mixin, unifying 4 delete expressions and 5 read predicates.

I brought the dashboard’s load from over 10 seconds to within its 5-second target by decoupling the source fetch and caching with React Query.

SOC 2 preparation: an audit of the backend API, and short-lived tokens with refresh. Later I found a mass-assignment flaw that let organisation members write billing fields and wrote the fix: an allowlist of editable fields and a partial unique index.

Jul 2022 – Sep 2024 2 yrs 2 mos Chennai Stack C#, .NET

Software Engineer at Quantitative Brokers

I built monitoring, risk and order-entry tools for a trading desk.

I designed a real-time order monitoring system that merges messages from several upstream sources into live aggregates and market-data ladders, so desk analysts see orders and market activity in one place.

In the risk-control application I moved data fetching from the client to the server, which took initial load from about 40 seconds to under 1 second.

measured
40 s before <1 s after

I optimised the critical execution path of a C#/.NET order-entry application, cutting latency by 20% and CPU usage by 50%, which made order processing faster and lighter on resources.

measured
−20% latency −50% CPU
On the order-entry application’s critical execution path.

I implemented DV01 risk calculations for the desk’s cash instruments: how much each instrument’s value moves for a one-basis-point change in interest rates.

Earlier roles

May – Jul 2021 Chennai Stack Angular, OpenFin, kdb+, Highcharts

Software Development Intern, Quantitative Brokers. I built a desktop integration prototype with OpenFin that shared context across four separate applications, and an Angular app that plotted financial curves from a live kdb+ database using Highcharts.

Feb – May 2021 Remote Stack Android, Firebase

Android App Developer, Prescribe. I built an Android app for collecting patient details and surfacing the feedback collected. It used Firebase for API endpoints and a custom notification system, and I instrumented it to analyse user interaction at each step.

May – Jul 2020 Remote

Software Development Intern, Workfence Technologies. I built a social-distancing mobile app paired with a hardware distance-measuring solution, and a geofencing-based attendance system with a visualisation interface that scales from 1,000 to 100,000 users.

Back to contents

Writing

29 Aug 2026 15 min read

Notes from building an SVG tool: what the numbers after C in a path actually mean, why fonts and CSS easing are the same curve, and why it is all one repeated lerp. With draggable playgrounds.

The only post so far. It came out of svg-reconstruction, listed above.

Back to contents

About

I work across the whole stack, from the schema to the interface, and the parts in between are the ones I want to own. I’m a founding engineer at Mando today. Before that I spent two years on real-time trading systems, where being slow was the same as being wrong.

I studied Chemical Engineering at IIT Madras, which taught me to think in systems, flows and balances long before I thought in APIs. Somewhere between a process simulation assignment and a hostel-room Android app, software won.

What I like is the part most teams hand off: a save from one tab silently overwriting another, a backfill that brought deleted docs back in production, a 4K video with roughly 8,000 annotations that hangs the editor. Those bugs live between layers, and the reliable way to fix them is to own both sides. Trading systems taught me the other half: a latency budget makes you honest, because either the number moved or it didn’t.

Outside work I build tools I want to use myself: leak, a terminal subscription manager in Go; deadfall, an npm CLI that finds unused code in Next.js, Vite and React projects; JumpKey, which turns URL templates into keywords; and a couple of Mac menu-bar apps. I follow cricket closely enough that a path-to-playoffs explorer for the IPL felt like a reasonable way to spend a season, and a multiplayer cricket card game followed.

Back to contents

Education and skills

2018 – 2022 Chennai

Indian Institute of Technology Madras

B.Tech. in Chemical Engineering, CGPA 8.58/10. JEE Advanced 2018: All India Rank 2,275.

Skills what I work with

Skills

Languages. TypeScript, Python, JavaScript, SQL

AI & LLM Engineering. LangChain, LangServe, Vercel AI SDK, RAG, Tool Calling, LLM Evaluation, Langfuse, LLM Streaming

Frontend. React, Next.js, Angular, TanStack Query, Zustand, Tailwind CSS, shadcn/ui, Storybook

Backend & Real-Time. Node.js, FastAPI, Fastify, WebSockets, Socket.IO, Pydantic

Browser & Media Engineering. Chrome Extensions (MV3), Chrome DevTools Protocol, FFmpeg, Video.js, Konva.js

Databases & Storage. PostgreSQL, Redis, Drizzle ORM, SQLAlchemy, Neon

Cloud & Infrastructure. GCP (Cloud Run, Cloud Build, Cloud Storage), Cloudflare (Workers, R2, Domains & DNS), Docker, Terraform, GitHub Actions

Testing & Observability. Playwright, Vitest, pytest, PostHog

Integrations. Stripe, Microsoft Graph, Nylas, Cloudflare, Recall, Knock

why No proficiency bars; nobody believes them anyway.

also, in personal projects Go, Swift, Astro, Svelte, three.js, Web Workers, SQLite.

Back to contents

Contact

Chennai IST, UTC+5:30

rishikeshpavithram@gmail.com

Email is fastest. If you’d rather talk, book a call.

There’s no form here; it would only have opened your mail app.

Back to contents