Chennai · est. 2004

Siddharth

AI/ML & backend engineer. I say yes to the unknown, then go all in.

← enter the voyage

Finding a Voice

2009–2021 — Abacus Montessori, Chennai

One school, start to finish. Where I learned to lead a room and argue a point.

House captain → Cultural Secretary

Vice captain, then captain of my house, then Cultural Secretary of the school.

A steady climb through school leadership — from vice captain to captain of my house, and finally Cultural Secretary of the school. Running events, wrangling people, and learning that organising is its own kind of building.

Debating

Deep in the debating circuit through all of school.

Debating was the through-line of my school years and carried straight into college. It's where I got comfortable thinking on my feet and making an argument land.

For the Love of the Problem

2021–2024 — Loyola College, Chennai

BSc Mathematics — chosen purely because I loved the subject. Also where I taught myself to build.

Why maths

Skipped the engineering/medicine track — studied maths because I actually liked it.

I picked maths out of genuine interest, not the default engineering-or-medicine route. Two years of academic scholarship, consistently among the top of the batch — missed the gold medal by a single spot.

Jet Sigma: 4 vs ~100 teams

Four of us against ~100 senior engineering teams. We won ~90% of the events.

At the Jet Sigma symposium (St. Joseph's College, ~2022–23), a team of four maths students went up against roughly a hundred engineering teams — more senior, more experienced — across a mix of events, and came out winning about 90% of them.

Teaching machines to play

Learned ML by building game-playing AIs instead of the usual tutorials.

I wanted to learn ML but skipped the clichéd starter projects — games were eye-catching enough to make me actually want to understand what was happening. Built a Flappy Bird agent with reinforcement learning, and a self-driving car (Trackmania-style) with Proximal Policy Optimization.

Judo detection @ IIT Madras

Pioneered a computer-vision system to track judo players — my first real applied AI.

A final-year sports-science internship at IIT Madras (with the Inspire Institute of Sport, Bellary). Coaches wanted to measure how long players spent in contact during throws, on a mat split into a grid of quadrants, each individually focusable from a coach's tablet. I was the first person assigned to it — built the initial infrastructure from zero, learned camera hardware from old low-level docs and computer vision from scratch, and left a foundation for the team to carry forward.

I'll Figure It Out

2024–2026 — Paperflite, Chennai

Walked in for a marketing role, said 'I'll figure it out' — walked out a founding AI engineer.

IDB document chatbot

A trust-first RAG chatbot for the Inter-American Development Bank — live and public.

For the IDB — one of the largest development-financing banks across the US and Latin America — trust was everything, so we had to nail RAG citations. It's fully live and public today. The bank's own researchers emailed to say how much easier it made researching their institution's stance on things.

Agentic analytics over 100M+ rows

Naive RAG wouldn't cut it — so I built an agent that writes and runs its own queries.

From the RFP phase with the IDB: a chatbot over survey datasets in the hundreds of millions of rows. Knowing naive RAG would fail at that scale, I built an agent-orchestration system with a code-execution tool that queries the database and runs Python analysis on demand. Delivered, and praised by the bank's researchers alongside the document chatbot.

Heysales call simulation

An AI that role-plays a prospect so salespeople can rehearse — it seeded a whole product.

Heysales was Paperflite's first LMS. I built the call-simulation feature — an AI prospect salespeople can practice against — shipping fast through the US shift, integrating live demo feedback and redeploying the same day. That initial pace seeded the entire LMS offering, which grew into the whole 'sales' side of Paperflite. It scaled from a pilot to 50–100 concurrent users.

Deal intelligence

Real-time prediction of how likely a rep is to close, from CRM + content signals.

An ongoing platform predicting, in real time, how well a salesperson will close their deals — from CRM metrics and how they share content and collateral. A continuous work-in-progress the whole time; piloted to a few customers rather than fully shipped.

Cleverstory: recommender + web→mobile

A recommendation system, and auto-converting web pages to mobile with stats + heuristics.

Two projects on Cleverstory, Paperflite's interactive content builder: a recommendation system, and a statistics-and-heuristics engine that automatically converts a web page into a mobile layout — solving the 'mobile is an afterthought' problem that plagues web builders.

New Signal

2026 — Lucio (current)

Legal-tech's first AI hire — got in by solving their CTF and texting the product lead. Now building the next product from the ground up.

Hired via a CTF

Solved the challenge on their careers page, texted the lead, got the call.

Lucio (legal tech) had a Capture-the-Flag challenge on their careers page — an API scavenger hunt where each response gives the clue to the next, until the final one yields the reward. I solved it, texted the HR and product lead, and that was that. First AI-specific hire at the company.

Building the next product

On the team building a new product from zero — core frameworks, languages, architecture.

The first month was AI-focused; now I'm moving into backend and core development. The company is pivoting into a new product and I'm part of the team building it from the ground up, in on the core decisions — frameworks, languages, architecture.

side mission

The Long Way

2022–present — endurance sport

Started holding my mum's spare water bottles. Ended up racing triathletes just behind the Ironman pros.

Triathlon

Consistently on/near the Chennai podium — just behind professionally-trained Ironman athletes.

Got in when a friend offered a spare triathlon ticket and I said yes without knowing what I'd signed up for. Since then I've consistently finished on or near the podium in Chennai triathlons — right behind professionally-trained Ironman athletes, as an amateur.

200 & 400km brevets

Long-distance endurance cycling — BRMs of 200km and 400km.

Cycling started because my mum was into it before me and wanted a partner to carry her spare water bottles. It turned into BRMs — long endurance rides of 200km and 400km. The endurance part is meditative; I like that it's mostly a mental game.

side mission

Off the Clock

ongoing

The rest of me: mountains, waves, and a camera.

Himalayan treks

Several treks in the Himalayas.

I travel and trek a lot — a handful of Himalayan treks among them.

Surfing

One of my ongoing passions.

Surfing is one of the things I keep coming back to.

Photography

Slightly into it — usually pointed at the places I trek and surf.

Lightly into photography, mostly from the treks and travels.