my work

Adsiwzz

Staff Software Engineer, 2016 — 2023

I joined Adswizz straight out of the uni in the Research and Innovation team led by the CTO at the time. Focus was and still is on building *the next big thing* in audio advertising. Started out with building java spring services and custom ETL pipelines for our VIP customers. At that time things got mainly done using AWS EC2 instances and unfortunately for me, who only previously dealt with (very) small data and no concern whatsoever for performance, we were dealing with real big data. I learned a lot about distributed systems, performance and scalability. I was fortunate to have a very talented team around me and I learned a lot from them. Things changed since those early days and today the combination of Redash and AWS Athena (along with other major infra changes across the org) serves us well. Then AWS lambda also became a thing so we explored that avenue as well and even though serverless is (definitely) not a silver bullet it does have its place. Nowadays with the ubiquity of containers we use a mix of AWS ECS and kube for our workloads. Although I can't disclose everything I've done here are some of the highlights:

• Software Engineer. Built a system that was responsible for raising addressability. The system is still maintained today and was continuously evolved and changed from its initial design to accommodate for current data privacy policies. Researched, implemented and wrote two patents. See below.
• In 2018 , I was promoted to Senior Software Engineer. Built and validated a proof of concept system for retargeting listeners. Today it lives in prod. Researched, implemented and wrote two patents. See below.
• In 2020 , I was promoted to Staff Software Engineer. Let's just say AI, LLMs and NLP -stuff and stuff.

Universitat Autonoma de Barcelona

PhD Candidate/ Marie Curie Early Stage Researcher, 2013 — 2016

I was a part of the Marie Curie Initial Training Network -TRANSACT where I researched brain cancer using a modality called Magnetic Resonance Spectroscopy. I was on the signal processing and machine learning side of things; cleaning up the data, extracting features and building supervised and unsupervised models. I focused primarily on matrix factorisation algorithms [1] applied to spectroscopic data. If you're interested in the topic you can see me present some of my work here and if you're really really interested drop me a line and we can chat.

Publications and Patents

Publications

• Automated Quality Control for Proton Magnetic Resonance Spectroscopy Data Using Convex Non-negative Matrix Factorization
• A machine learning pipeline for supporting differentiation of glioblastomas from single brain metastases
• From raw data to data-analysis for magnetic resonance spectroscopy – the missing link: jMRUI2XML

Patents

• Identifying personal characteristics using sensor-gathered data
• Determination of user perspicaciousness during a content stream
• Delivering tailored audio segments within live audio streams
• Decoupled custom event system based on ephemeral tokens for enabling secure custom services on a digital audio stream