cv

Engineering Director of Machine Learning

Adevinta

Leading the central Machine Learning Tribe of Adevinta, with the mission to make machine learning very easy and impactful for the rest of the teams in the company, with high focus on Generative AI.

Invited teacher / learning facilitator on machine learning

Different universities

1) Invited teacher on machine learning at ESADE Insurtech online training. 2) Providing seminars, content, exercises, individual sessions, etc. for other online courses in ML in two of top 10 US universities. 3) Apr 2018 - Feb 2021 · 2 yrs 11 mos Barcelona, Catalonia, SpainBarcelona, Catalonia, Spain Providing a quarterly seminar on how machine learning is done at scale in the enterprise world to the students of the Economics and Business Faculty of Universitat Pompeu i Fabra.

Head of Product of Machine Learning Services Area

Adevinta

Co-heading the Machine Learning Services area of Adevinta. As Product Lead, I am responsible of "what the area does": objectives, products, relationship with stakeholders and strategy. The area is composed by three teams, and serves all our marketplaces globally. Our mission: "make machine learning easy and impactful to all teams in Adevinta". During 2020 our company was able to deliver more ML than in the previous 4 years, totalling more than one hundred ML use cases. The number of teams doing ML increased by 60%, the time to build and deploy ML models was reduced by 15%, and hundreds of people were trained in ML. Some of the main initiatives of our area: (1) We improve Adevinta marketplaces with our ML models: from image recognition to recommender systems and NLP. The results of our models impact millions of users daily, helping them find what they are looking for, making the marketplaces safer, or reducing our costs and increasing our revenue. (2) We nurture an active internal ML community with hundreds of members to share learnings, best practices, tooling, use cases, etc. (3) We build tooling to make ML teams more effective, by working very close to the ML community. During 2020 we were able to build and deploy an ML platform to 80% of teams in Adevinta, providing them with new capabilities and reducing their time to build and deploy ML by 15%. (4) We increased the knowledge and adoption of ML with in house online trainings for both business and tech people: the ML academy. Hundreds of participants finished it, rating over 92%, and more than half of the participants identified new opportunities with ML in their business domain.

Co-writer of textbook - Machine Learning and Big Data

IE Business School

Co-authored the Machine Learning and Big Data section of the "Manual of Economic and Competitive Intelligence", which is used as main textbook of the Economic Intelligence courses of Instituto de Empresa, Universidad Autonoma de Madrid and Universidad Rey Juan Carlos. Book was published by editorial Tirant Lo Blanch, ranked as best academic editorial in Spain.

Senior Data Scientist

Schibsted Media Group

Collaborating with over 10 teams in the company, delivering value on models related to SEO, SEM, UX, Security, working with over 8 marketplaces. Built and deployed models with different levels of complexities, from deep learning (specially NLP) to more traditional machine learning (e.g. gradient boosting). Normally worked embedded for a few quarters inside engineering teams. Used to work with agile methodologies (Scrum) and collaborative tools (Jira, Github...). Some of the tools & algos frequently used: TensorFlow, Pytorch, Scikit Learn, Pandas, GPU model optimisation, TensorBoard, Spark, Docker, AWS, Linux, Python, Scala, Jupyter. Advising management on scaling machine learning across the company and marketplaces. Advice given in form of memos, presentations or informal discussions. One of the major barriers to scale machine learning is the knowledge gap. Creator of internal workshops "Machine Learning and Deep Learning for coders" & "ML for Business", over 200 people trained in several 1 to 3 days sessions, rated 4.8 out of 5 (105 raters).

Machine learning business and technical consultant

Self-Employed

Worked for (1) a multi-awarded hedge fund, (2) the largest asset manager of Europe and (3) a FinTech startup. Projects included: - Interim CEO of a $5M media company specialized in automatic trading during transition after buyout. We were able to make it profitable again. - Developing of quantitative strategies and solutions to trading desks. - Developing BI tools that allowed sales teams discover new opportunities of the order of hundred million USD notional.

Senior Data Scientist, Quantitative Analyst

Morgan Stanley

Worked in the front office equities trading desk: - Developed, maintained, and help selling/marketing automatic investment strategies which brought more than $1.5 billion of assets under management. - Optimum portfolio construction and analysis using improved estimators of the covariance matrix. Employed to reduce volatility, hedge assets, increase liquidity, etc. - Morgan Stanley was awarded with several prizes by independent bodies which highlighted quantitative strategies which me and my team developed. The awards included Best UK Structured Product House (2013) and Best Equity Derivatives House (2013) by Structured Products magazine and International Financing Review. See, for example http://goo.gl/2oASZi and http://goo.gl/tx6dBn

Product Manager

Arbora & Ausonia (Procter & Gamble JV)

A&A was an Iberian Consumer Goods company, which belonged to Procter & Gamble. The company was the market leader in more than six multi million euros brands. Business Unit Product Manager (2006-2009) - Responsible for launching new products to the market, leading a multifunctional team of 11 members. - Increased top brand profit double digit vs previous year, using best in class P&G marketing framework. Production Planning Manager (2003-2006) - In charge of three members team, responsible for planning the production of three production plants. - Coordinator of international teams (Europe, Japan & USA) for Import/Export projects.

Mentoring

Harris Academy Bermondsay

Mentoring of students. I look to inspire and help them by communicating positive values and speak about the implications of their own career decisions.

MSc

Imperial College London

Mathematics and Finance

Postgrade

Universitat Pompeu Fabra

Marketing

MSc

Imperial College London

Theoretical Physics

MSc

Universidad de La Laguna

Physics

Gold medal in Kaggle Data Science competition (position 11 out of 4,037)

Kaggle

I used a Deep Learning NLP model implemented with PyTorch and torchtext, in the Quora Insincere Questions Classification (detect toxic content to improve online conversations). https://www.kaggle.com/manuelsh

Best UK Structured Products House, Best Equity Derivatives House

Structured Products magazine, International Financing Review

Several awards won by Morgan Stanley highlight quantitative strategies under my development. http://goo.gl/SEkDKf http://goo.gl/2oASZi http://goo.gl/ehndQI

Communication Awards

As product manager in P&G, awarded best spot in international advertisement festival (2010), Best Social Campaign 2008 (internal P&G prize), Best Press Campaign given by national magazine (2008), Best Advertisement of P&G Marketing Communication Seminar (2007).

Scholarship Caja Madrid

Caja Madrid

Awarded with a scholarship from Caja Madrid which fully sponsored the MSc in Mathematics and Finance, including living expenses

Collaboration Scholarship

Spanish Ministry of Education

Awarded with a scholarship which allowed me to collaborate with the Fundamental Physics Department on a research which culminated later on a publication on Physics Review.

National Physics Olympics

Fourth prize in the Spanish National Physics Olympics.

Escape of photons from two fixed extreme Reissner-Nordström black holes

Physics Review

We study the scattering of light (null geodesics) by two fixed extreme Reissner-Nordstr\"om black holes, in which the gravitational attraction of their masses is exactly balanced with the electrostatic repulsion of their charges, allowing a static spacetime. We identify the set of unstable periodic orbits that constitute the fractal repeller that completely describes the chaotic escape dynamics of photons. In the framework of periodic orbit theory, the analysis of the linear stability of the unstable periodic orbits is used to obtain the main quantities of chaos that characterize the escape dynamics of the photons scattered by the black holes. In particular, the escape rate which is compared with the result obtained from numerical simulations that consider statistical ensembles of photons. We also analyze the dynamics of photons in the proximity of a perturbed black hole and give an analytical estimation for the escape rate in this system.

Portfolio of Best Systematic Investment Strategies

After exploring and backtesting hundreds of ideas, has constructed a portfolio of the best ones, which include statistical arbitrage and effective alpha extraction. Most of the ideas are uncorrelated to the market, have a low exposure time (sometimes less than half an hour a day) with no overnight risk, and have high Sharpe Ratio.

Contributor to Transformers Library

Added dynamic resolution input for all Clip family models in the Hugging Face Transformers library, improving the following models: altclip, bridgetower, chineseclip, clip, clipseg, kosmos_2, x_clip and git. Listed as "Significan Community Contributions" in the release 4.46.0. https://github.com/huggingface/transformers/releases/tag/v4.46.0

On the Optimization of a Fund of Funds

Download project on http://goo.gl/53SJpi Grade A. With Morgan Stanley’s Quantitative & Derivatives Strategies group, used real market data of more than 3, 000 mutual funds to build optimum portfolios. Implemented and backtested different optimization models: Markowitz, Shrinkage, Multifactor, PCA… After the project I was offered a permanent position at Morgan Stanley.

Personal Portfolio Manager

Following value investing principles have outperformed the market during the whole investing period: +14.3% vs market in 2010 +11.2% vs market in 2009 Books recommended: - The Essays of Warren Buffett: Lessons for Investors and Managers - Warren Buffett and the Interpretation of Financial Statements: The Search for the Company with a Durable Competitive Advantage - The Intelligent Investor. Benjamin Graham - Security Analisys. Benjamin Graham and Frank Dodd

Senior Investment Strategist at the Imperial College Investment Club

Developed the long term strategy for an active management fund for students, with a value investing approach.