Raghav Jerath

Raghav JerathMay 18, 2026

Paris → Singapore → New Orleans And now London. My next chapter is now locked in. At 17, I moved a few thousand kilometres for ESSEC Business School. No real playbook for it. Figured things out as I went, how to learn a new language, how to walk into unfamiliar rooms and find my footing, among many. Four years, three continents. Ended up working at startups to grasp as much. Played with many AI tools in Product and GTM, learning alongside exceptional co-founders and talent what it takes to help early-stage startups scale with momentum and traction. Also spent the last few months researching Retail AI, as a part of my BBA thesis (more on that soon). Somewhere in all of that, I built a lot more clarity. AI inside an org is never really a tech problem. It's usually a people problem, a process problem, and a change problem. The tools are often the easy part, but the environment around them isn't. That's the gap I want to get sharper at. And the 𝐌𝐒𝐜 𝐢𝐧 𝐃𝐢𝐠𝐢𝐭𝐚𝐥 𝐓𝐫𝐚𝐧𝐬𝐟𝐨𝐫𝐦𝐚𝐭𝐢𝐨𝐧 𝐌𝐠𝐦𝐭 at ESCP Business School is built around exactly that. A tech-specialized master's in one of Europe's leading B-schools. London & Paris-based. French b-school roots, continuing. Couldn't be more glad to share this! A special thanks to Alex Issa, PhD, Robin Philibert, and Fady B., without whom I wouldn't have been here. ESCP Business School , see you soon. 👋

Raghav JerathMay 21, 2026

I trained an 𝗠𝗟 𝗺𝗼𝗱𝗲𝗹 on 𝟰𝟵,𝟮𝟭𝟱 football matches to predict the entire Fifa World Cup 2026. Here's who wins. I've been building many projects lately, and this one is something I’ve been very passionate about. The model is built on 300 decision trees trained sequentially, each one correcting the mistakes of the last (a technique called Gradient Boosting) It uses: • ELO ratings (the chess ranking system, adapted for football) • Recent form (last 5 games) • Goals scored per game • Home advantage It has two modes: 𝐇𝐞𝐚𝐝-𝐭𝐨-𝐇𝐞𝐚𝐝: pick any two World Cup nations, choose neutral ground or home advantage, and get win/draw/loss probabilities with these ELO ratings and recent form side by side. 𝐅𝐢𝐟𝐚 𝐖𝐂 𝟮𝟬𝟮𝟲: runs the full simulation, all 48 teams, 12 groups, the official bracket, 104 matches, and predicts the brackets and the winner. Here’s a few things worth knowing: • The 49,215 matches are used to compute ELO ratings across 150 years of football. This model itself is trained only on matches from 2014 onwards; the modern game is far too different from earlier eras. • Cross-validated accuracy sits at ~ 58% (the model can be incorrect 42% of the time) • The group stage uses expected points (fractional), not discrete wins, which makes rankings more stable. Tech Stack: Python, scikit-learn, Flask, deployed on Vercel. Frontend: Vanilla HTML / CSS / JS. --- Link to the live project in the comments. 👇 Try it yourself and comment which prediction you disagree with most.

Raghav JerathOct 8, 2026

Anyone else feel like London runs at a different speed right now? 🇬🇧 Life update: I've just moved here for the next 6 months for a Master's in Digital Transformation at ESCP. The weather welcomed me with sunshine, then reminded me where I was. Fair enough. What I didn't expect was the energy. London startups raised $14bn in the first half of 2026. That's 28% of all venture funding across Europe, the Middle East and Africa, the city's highest share on record. And you feel it on the ground. Founders pitching in coffee queues. Investors at every event. Engineers building AI products on the Tube home. Honestly, it rivals SF. The talent. The infrastructure. The universities. The capital. It's contagious. And I'm all in. So I'm stepping up as President of the ESCP AI Society! A community where students, founders, investors, and builders actually meet, share ideas, and build together. More talks with people in the AI space. More connections with London's startup and VC ecosystem. If you're in London and working in startups, VC, AI, or anything close, let's connect. I'd love to hear what you're building, and if you'd like to speak at or partner with the AI Society, my inbox is open. 🚀 Let's do this, London!