Max Grigoryev

Product @ LSEG — Analytics & AI

Endurance athlete. Builder. Long-distance thinker.

I simplify complex financial systems, build tools for developers and AI models, and spend a lot of time training for long races.

About Me

I build products at the intersection of AI, analytics, and developer tooling. Right now: leading product for LSEG Analytics & AI, shaping how financial markets are understood by humans and LLMs.

Outside of work, I'm deep into endurance sports — 3× Ironman 70.3 finisher (Valencia, Berlin, Venice), full Ironman finisher in the Loire Valley, and 4× marathon finisher with a 3:07 PB. Currently training for Ironman 70.3 Cascais.

The same obsession shows up in what I build after hours: metrica.fit and Onda, two iOS apps for endurance athletes; beetroute, a free route planner that also ships inside Onda; and chiptime, an open-source library that rescues corrupted workout files.

I like long distances — in products, in training, and in life.

Work @ LSEG

Lead product management for LSEG Analytics & AI

AI for Financial Markets

Building the semantics, ontology, and tooling that let LLMs interpret pricing, analytics, curves, events, and research at scale.

Developer Platforms

APIs, SDKs, runtimes, abstractions — the infrastructure that makes financial data feel clean, fast, and consistent.

Simple Philosophy

Complexity deserves clarity. Products should feel inevitable once they exist.

Endurance & Triathlon

Training is where I reset, rethink, and rebuild. The process is hard, and that's the point — it's the discipline it builds that carries into everything else.

"Only the disciplined ones are free. If you are undisciplined, you are a slave to your moods"

Eliud Kipchoge, world's fastest marathoner
NEXT UPIronman 70.3 Cascais → and whatever the next start line turns out to be

How I Train

Discipline. Passion. Motivation. In that order — because motivation fades, passion is the why, and discipline is the must.

Not a Pro… Technically

I'm not a professional athlete — I just play the part with enthusiasm. My quads are still questioning the casting.

Find the Data

A quiet record of the miles, the early mornings, and the long sessions that shape how I work — patient, structured, data-driven.

Metrica.fit

AI endurance coaching for triathletes and runners — a full season of training that re-plans itself every week.

The Problem

Endurance plans come in two shapes: a static PDF that dies the first week life interferes, or a coach most athletes can't justify. Everyone else guesses.

What It Does

Builds a season around your actual races, then re-optimizes every week against what you completed and how you recovered. Data arrives on its own from Apple Health, Strava, Garmin, and WHOOP.

Free, Where It Counts

ML bike-fit analysis from a phone video, IRONMAN and 70.3 race guides, and a World Championship slot tracker stay free — the parts of the sport nobody should have to pay to understand.

Keep athletes focused on what matters, and give something back to the sport that changed me.

Onda

An endurance journal for iOS — where a ride, run, or swim goes after the watch stops. Free, chronological, human.

The Problem

The places athletes park their training have drifted into engagement mechanics — algorithmic feeds, streak nudges, a paywall prompt one mis-tap away. The people who train hardest stay for the social graph, not the product.

What It Does

Nothing to press start on. Garmin, Suunto, and Wahoo sync on their own, Apple Health brings in the rest, and Zwift emails land at a personal inbox address. Then the map becomes the hero: splits, power curves, best efforts, training load, and share cards worth posting.

Free, and Human

No subscription, no ads, no badges or streaks. The feed is chronological and every notification comes from a real person. Clubs get their own page, feed, and leaderboards.

Your training should look as good as it felt — and the record of it belongs to you.

Beetroute

A free route planner for athletes — no account, no subscription, just plan and ride. The same planner ships natively inside the Onda iOS app.

The Problem

Every good route planner asks for a subscription before it will draw a line — and none of them know what session you're actually meant to be doing today.

What It Does

Self-hosted routing over OpenStreetMap with road, gravel, and run costing. Loops generated from wherever you're standing, real elevation profiles, and GPX or FIT export straight to your head unit.

Workout-Aware

Give it the session — say 4×8 minutes at threshold — and it looks for terrain that fits the intervals, instead of handing you a climb in the middle of every rep.

Routing infrastructure is cheap to run and expensive to buy. That gap is the whole product — on the web, and in your pocket.

chiptime

Recovery-grade FIT file processing — parse anything, lose nothing silently, explain everything.

The Problem

Watches die mid-ride, batteries quit, firmware writes nonsense. Every open-source FIT parser stops at the first bad byte, and the tools that can actually repair a file are closed GUIs you can't automate.

What It Does

Resynchronizes mid-file, salvages truncated recordings, rebuilds missing session summaries from the raw records — then writes a valid .fit back out that platforms will accept.

Nothing Silent

Every drop, repair, and reinterpretation lands in a machine-readable trail. Sentinels become null before any statistic touches them, so a dead sensor never quietly reads as zero.

MIT licensed, zero runtime dependencies, Python 3.11+. The edge cases are the product.

Contact