pere / portfolio
whoamiBased in Europe

Pere Manresa

Product Data Scientist
/ AI Engineer

Pixel-art portrait of Pere Manresa in warm orange tones

Experience

2017—present
2022 → nowRemote, Europe

Shopify

Product Data Scientist / AI Engineer

  • AI that does its homework

    I build analytics agents that combine Shopify’s business context with its data warehouse to help teams answer their own questions. RAG, data freshness checks, automated evaluations in CI/CD, and production monitoring keep the answers grounded.

  • A faster trip through checkout

    I helped build POS features that connect customers with merchants across Shopify’s network, improving customer identification and marketing opt-in. I also contributed to checkout latency improvements of more than 30%.

  • Product questions, data answers

    Across checkout, buyer identity, customer data, and retail, I turn open-ended questions into evidence that shapes roadmaps, launches, and measurement for products serving Shopify’s 2M+ merchants.

  • The foundations teams build on

    I’ve built 100+ reusable data models and 50+ dashboards, alongside instrumentation plans and KPI frameworks. Product and engineering teams use them to track performance and decide what to ship next.

  • Python
  • SQL
  • AI agents
  • RAG
  • Prompt engineering
  • Evaluations
  • CI/CD
  • Product analytics
  • Data modeling
  • Instrumentation
2019 → 2022Berlin, Germany / Remote

BCG Digital Ventures

Senior Data Scientist / ML Engineer

Joined as a Data Scientist in 2019; promoted to Senior in July 2021.

  • Six products, very different puzzles

    I worked inside client and venture teams to build, launch, and scale six products: car warranty automation, a vision care recommender, tissue production optimization, construction bidding, marketing segmentation, and a consumer scent curation engine. Yes, even scent.

  • Models all the way to production

    I built ML pipelines covering data collection, feature engineering, model development, deployment, and monitoring. I also worked with the teams using them to make the models part of the actual product.

  • A varied ML toolkit

    The work included recommender systems, multi-armed bandits, NLP for taxonomies, process optimization, and customer segmentation. I helped teams use the results to make product and business decisions.

  • Growing the team

    I led hiring, onboarding, and mentoring for junior data scientists, alongside the hands-on engineering work.

  • Machine learning
  • Feature engineering
  • ML pipelines
  • MLOps
  • Recommender systems
  • NLP
  • Multi-armed bandits
  • Customer segmentation
  • Product development
2017 → 2019Düsseldorf, Germany

trivago

Data Scientist / ML Engineer

  • ML at travel-search scale

    I trained, deployed, monitored, and scaled ML models on Hadoop and Spark infrastructure serving millions of users. The work improved cross-channel marketing ROI by more than 30%.

  • Less waiting on Spark

    I optimized an Oozie/Spark ETL workflow, cutting runtime from five hours to about 45 minutes. A much shorter wait for the next result.

  • Making the work useful

    I built KPI dashboards and analytical tools for senior stakeholders, and shared what we learned through internal ML workshops, meetups, and panel discussions.

  • Apache Spark
  • Hadoop
  • Oozie
  • ETL
  • Production ML
  • Model monitoring
  • Marketing analytics
  • Data visualization

Side projects

4 projects / Built outside work
trainhybrid.app
Hybrid landing page: Train like your disciplines are connected
02

TrainHybrid

in development

Smarter training for hybrid athletes who lift and train for endurance.

About & stack

I’m building it to train better and smarter, with strength and endurance planned together around the needs of hybrid athletes.

  • JavaScript
  • HTML / CSS
  • Cloudflare Workers
  • Swift / SwiftUI · in progress
  • AI coaching · planned
  • HealthKit / Strava · planned
Visit trainhybrid.app ↗
sheliefimova.com
Sheli Efimova portfolio: Global Creative Production Lead over a video reel
03

Sheli Efimova

live

My wife’s creative production portfolio, with campaign reels and full-length films.

About & stack

My wife’s portfolio. I built a place for her campaign films and creative production work to speak for themselves.

  • Astro
  • Tailwind CSS
  • Python
  • FFmpeg
  • Cloudflare Workers
  • R2
  • GitHub Actions
Visit sheliefimova.com ↗
espicotcm.com
Es Picot run club homepage with a trail runner photo
04

Es Picot

live

A website for my village’s running club in Felanitx, Mallorca.

About & stack

A fun collaboration to give my village’s running club a home on the web, with memberships and race sign-ups.

  • TypeScript
  • React
  • React Router
  • D1 / SQL
  • Cloudflare Workers
  • Mollie
  • Clerk
Visit espicotcm.com ↗

A bit about me

Away from the day job

I love lifting
and endurance training.
Always finding
things to build.

I like being the bridge between product and technology: understanding what someone needs, then figuring out how to build it.

AI is a big part of that curiosity. I’m interested in the path toward AGI and what increasingly capable systems will let us make. I want to understand the technology well enough to put it to work.

Outside work, I love lifting, endurance training, and triathlons. I’m currently training for HYROX, putting strength and endurance to work together.

2016 → 2017

MSc, Fundamental Principles of AI & Data Science

Universitat de Barcelona
2012 → 2016

BSc, Computer Science & Mathematics

Universitat de Barcelona

Earlier work

Research & open source

Get in touch

Say hello.

GitHub ↗LinkedIn ↗X ↗
›esc