48.8566° N — 2.3522° E
Paris · --:--:--
Open to opportunities
Engineer × Builder

Aditya Singh

Engineerby trainingFinanceby choiceBuilderby default

Paris, Île-de-France · Expedition in progress

Scroll to explore

Who's climbing

The short version

I started as an Engineer and spent 3 years deep in data engineering, and somewhere along the way I started two companies, both of which failed, and taught me more than anything that worked.

I like being the person who'll figure out whatever needs figuring out, whether that's a broken pipeline, a GTM problem, or an AI agent I'm building at 1am just to see if I can.

Now I'm at ESSEC (Grande École MiM, Finance major), learning and experiencing first hand on how AI is changing the Finance world.

Logbook records

  • Global Circular Economy Chair — 1st Prize, awarded by the French Ministry of Ecological Transition, 2026
  • 1st of 3,000 — Niyoto Catalyst national case & analytics competition, 2023
  • Gold Medal, State Level — International Mathematics Olympiad (SOF), 2021
  • Microsoft for Startups Founders Hub — $350k Azure / OpenAI credits
  • AWS Certified Cloud Practitioner · Google Data Analytics

The route so far

Expedition log

Every stop taught something the next one needed. Scroll to climb.

UNNAMED PEAK

ALT 3108Ft — BASE CAMP

Dayananda Sagar College of Engineering, Bengaluru

B.E. Electronics & Communication

2019 — 2023

The engineering foundation everything else is built on — CGPA 8.4/10.

  • Google Developer Student Clubs Lead — ran cloud, data, and dev workshops for the campus
  • Captained the college football and basketball teams for two years

ALT 1 4000Ft — CAMP I

Alcon Laboratories

Data Engineer — EMEA & APAC

Jul 2023 — Jul 2025 · Bengaluru

Two years of production data engineering: 16 pipelines on Azure and AWS (Databricks, Glue, Data Factory) in Python, Scala, and SQL — processing $9bn in annual transaction volume.

  • Improved forecasting accuracy 35% and cut operational costs 20% with ML and anomaly-detection models
  • Reduced quarter-end closing by 3 days and saved 15+ hours/week through automation
  • Established data governance across Finance, IT, and Sales, partnering directly with non-technical users from requirements to adoption

SIDE EXPEDITION — The founder trail

Jharokha

Founder

2024 · Rajasthan, India

A hand-block-printing textile brand I started in rural Rajasthan from $400 of personal savings — grown to $12k ARR in six months before funding ran out.

  • 8 product lines, 2,500+ customers, and a manufacturing line built with 6 underprivileged but exceptionally skilled artisans
  • Funded the artisans' children's schooling and their families' health insurance which is still going on even after the brand wound down

SIDE EXPEDITION — The operator trail

Legend of Toys

Founder's Office — first hire

3 months · Bengaluru

Joined as the first hire alongside three founders — a Founder's Office intern with a hand in nearly every part of the business.

  • Worked across the production line, sales, marketing, product testing, and product photography
  • Drove fundraising and investor relations, and ran online sales on Amazon & Zepto

ALT 3 600M — CAMP II

ESSEC Business School

Grande École MiM — Finance major

Aug 2025 — 2027 (expected) · Paris

Building the finance layer on top of the engineering foundation — CGPA 16.8/20. Corporate finance, valuation, financial markets, strategy.

  • Portfolio Management Analyst, Economics Club (ECE), running a virtual €100k equity portfolio with DCF and comparable analysis
  • Member, Global Circular Economy Chair

ALT 4 700M — CAMP III

Equans (Bouygues Group)

Consultant — Sustainability, Data & AI

Feb — Jun 2026 · Paris

Built and piloted an LLM-powered tool with the HVAC procurement team that turned unstructured supplier documents and EPDs into structured material data — the project that won first prize from the ESSEC Global Circular Economy Chair, presented by the French Minister for Ecological Transition.

  • ≈35% time saved and 98% extraction accuracy vs. manual Excel entry, leading the team to weight supplier sustainability practices more heavily in contracts
  • Modelled exposure and risk across a 170,000-supplier base; benchmarked Michelin, Legrand, Saint-Gobain, Nexans, and SNCF against market indicators
  • Built KPI and monitoring frameworks; presented to senior leadership including the Group CSO

SUMMIT — UNNAMED PEAK

The third venture. AI × finance. Currently gathering supplies.

Built at basecamp

Things I've shipped

Fewer projects, deeper stories. Each one answers three questions: what problem, what I built, what it taught me.

AI · MARKETS

Morning Edge

PROBLEM
The Head of Strategic Innovation at BNP Paribas CIB Global Markets needed daily market intelligence distilled into a briefing that was actually readable — not another dashboard.
BUILT
A Claude-powered market-briefing tool built for that single demanding user: I designed the system prompts, the structured output formats, and the edge-case handling that keeps it reliable when markets get weird.
LEARNED
Building for one senior user with zero tolerance for noise teaches you more about prompt architecture than any tutorial — every word in the output has to earn its place.
Claude APIPrompt engineeringPython
View project →

AI · SUSTAINABLE FINANCE

GreenCheck

PROBLEM
Green bonds, sustainability-linked instruments, and Article 8/9 funds make claims that are hard to verify — and greenwashing hides in the gap between the claim and the evidence.
BUILT
A Claude-based claims-substantiation engine that reviews the instruments and flags each claim as substantiated, partial, or unsubstantiated.
LEARNED
This is the intersection I want to build in — AI applied to sustainable finance. It's the through-line from my Equans work to where the route leads next.
Claude APIRAGPython
View project →

AI · AGENTS

Voice AI Agent

PROBLEM
I wanted to understand how LLM agents actually reason — and where they fail — by building one from scratch instead of reading about it.
BUILT
An agentic voice receptionist: OpenAI API, RAG pipelines, tool-calling, and memory, wired together end-to-end with no framework doing the thinking for me.
LEARNED
Instruction design shapes output more than model choice. Watching an agent fail in production-like conditions is the fastest way to learn what robustness means.
OpenAI APIRAGTool-callingPython
View project →

AI · WEBGL · NEXT.JS

This Website

PROBLEM
A CV tells you what someone did. It doesn't tell you who they are — and it definitely can't answer your questions.
BUILT
This site: WebGL night sky, a scroll-driven expedition route, and an AI guide powered by Claude that answers questions about me in text and in my own cloned voice.
LEARNED
Shipping beats planning. Every effect here earned its place or got cut.
Next.jsTypeScriptWebGLClaude APIElevenLabs
Everything else lives on GitHub →

Off the clock

Beyond the grid

WHY THE THEME ISN'T A METAPHOR

Out of my depth, on purpose

Sixteen treks across India — the Himalayas, the Nilgiris, the Western Ghats. Beas Kund, Hampta Pass, Rupin Pass, Laka Glacier, Kudremukh, Triund, and more. This winter I start in the European Alps. The dream: the highest peak on every continent. It's the same instinct that makes me start building things before I know how they'll work — I like being somewhere I don't fully know how to handle yet, and figuring it out.

NATIONAL & STATE LEVEL

The athlete

Badminton at national level; football, cricket, and basketball at state level — and I captained my college football and basketball teams for two years. I still train six days a week, built around sprint sessions.

WHAT I READ & WATCH

Corporate power & financial crime

I'm drawn to morally complex stories about how business actually works — Succession, The Big Short, the Madoff saga. Not for the spectacle: for the mechanics of how systems bend and break.