About
Welcome to my own little corner of the internet.
I'm a principal data scientist working at a large Swiss bank. I created this space as a sort of "living resume", with a bit more color and character than a standard A4 page provides. I'll try to keep it current.
Thanks for reading!
Quick Facts
Years Data Experience academics and industry
Swiss 4000-ers summited by both foot and ski
Largest team led XENON1T data analysis team
Different countries where I've presented talks or lectures
Programming languages in which I've delivered productive projects
Countries I've lived in for at least 5 years
Years working in German as the official company language
Horsepower in my souped-up micro-hatchback
Resume
Please get in touch for an up-to-date version of my resume. A general version appears here below.
Sumary
Dr. Daniel Coderre
Experienced data expert with over a decade of experience in all aspects of the data project stack. As an emergent leader, I take ownership of projects and drive them to completion. Comfortable in both the server room and the board room, I like to contribute both to the high-level strategic planning and the nitty-gritty details.
- Bern, Switzerland
- daniel@coderre.ch
- linkedin.com/in/daniel-coderre
Education
Doctor of Natural Science, Physics
2012
Ruhr University Bochum, Germany
PhD work studying rare decays of the eta meson by accelerating protons at the COSY synchrotron, colliding them with protons/deuterons, and measuring the byproducts with the WASA detector.
Master's Degree in Physics
2009
State University of New York at Albany, NY, USA
With a major in physics. Thesis topic was training neural networks on high-dimensional, sparse input data to try to learn the detector response to simulated ATLAS trigger menus. Localized backgrounds were generated in populated areas of the parameter space using Markov Chains.
Bachelor's Degree in Physics
2007
State University of New York at Albany, NY, USA
Major in physics and minor in mathematics
High School Diploma
2003
Scotia, NY, USA
Professional Experience
Principal Data Scientist
2020 - Present
PostFinance, Bern, Switzerland
- Constructed a modern, open source AI platform from scratch based on the python data stack. Provides feature extraction and automated model training, integrated in a highly regulated enterprise application landscape
- Built an automated system using machine learning to intelligently reduce money laundering alerts saving up to a million francs a year in manual review costs
- Developed and maintained automated machine learning models to generate leads for marketing activities across numerous products in retail banking, includes product affinity, churn, and behavioral prediction
- Installed and maintained various infrastructure components, frameworks like Spark, Airflow and MLFlow, Postgres and MongoDB databases, productive web dashboards (Angular JS) and API systems (flask). Integrated everything into a strictly regulated enterprise IT landscape.
Postdoctoral Researcher, XENON Dark Matter Search
2017 - 2020
University of Freiburg, Germany
- Data analysis group coordinator: led an international team of over 60 scientists to produce the world’s most sensitive dark matter result (Read the Paper)
- Data acquisition group leader: led a team of 10 to design and install the data acquisition system for the XENONnT detector upgrade (Read the Paper)
- Data analysis and analytical software development. Design and implementation of hardware and software components for the data and analytics program
- Advised master and PhD students locally, taught physics, evangelized our results in numerous international conferences
Postdoctoral Researcher, XENON Dark Matter Search
2013 - 2017
University of Bern, Switzerland
- Designed, installed, and supported data acquisition system for a large-scale dark-matter search producing over 1PB of data per year during operation and with nearly 100% uptime, day and night, over years of data-taking (Read the Paper)
- Extensive work in analytical software development and data analysis, including leadership of an analysis sub-team tasked with investigating operational anomalies
- Teaching, mentoring of students, international talks, and outreach
PhD Researcher, WASA-at-COSY Experiment
2009 - 2013
Jülich Research Center, Germany
- Data analysis and development of scientific software (C++) probing rare light meson decays at the WASA-at-COSY Experiment (Read the paper)
- Placed upper limit on a possible non-standard-model CP-violating contribution to a specific decay of the eta meson
- Occurs to me now I have a penchant for trying to measure things that aren't there
Technologies I use Today
Technologies come and go. This is what my stack looks like today, but 5 years ago it looked a lot different and 5 years from now will be different as well.
Python, C++, Javascript, Golang
Everything in ML is python now but I have a soft spot for C++ and an academic interest in shoehorning go in where I can. I've maintained several web portals in the past and have built things in vanilla JS, react, and angular.
Analytics and Machine Learning
I cut my teeth working for years analyzing particle physics data in C++. But now it's all in python. My current stack includes tensorflow, scikit-learn, xgboost, prophet, pandas, numpy, hyperopt, and so on. Python is where it's at for anything data and ML at the moment and I've been working in this ecosystem for about a decade.
Cloud / Container
The genie is out of the bottle now and we're never going back to non-containerized applications. Right now I'm heavy into Kubernetes and work predominantly on-prem (banking...). But in the past I've build cloud-based systems as well, and I'm eager to return to the flexibility that the public cloud provides.
Data Engineering
I've been hands on with spark basically daily for about 5 years now. I'm not talking about databricks either, we do it on hard mode with pure on-prem spark. This is the tool with by far the steepest learning curve, but it's so powerful that it's worth it in spades. I also do a lot of work in Airflow and we've automated just about all our systems in this tool. SQL goes without saying.
DevOps / MLOps
Git, gitlab-CI, ArgoCD, helm, ML-Flow, splunk, prometheus, grafana... the number of tools used to create, deploy, monitor, and maintain a modern machine learning application can seem staggering at first. But these tools work so well together, it's more like learning different aspects of the same system. I pride myself in being one of those data scientists who can build a system from business problem back to business solution, including everything required along the way, and ML-Ops is a big part of this.
All the rest...
Stick me in a dusty lab in a University somewhere and I'll (happily!) build you a little partcle detector. I didn't forget the years I worked as a physicist doing a little bit of everything. Somewhere buried in my head is knowledge on NIM/VME electronics and signaling logic. Or writing network code in C using sockets. Or designing, configuring, and installing a server room, considering cooling via hot and cold aisles, local networking, grounding, and all those details. I like tackling new challenges, and having a broad experience from the past helps in doing so.
Photo Reel
A few pictures from work and life for the past several years.
XENONnT Amplifiers
Low-voltage signals are routed several meters, through a vacuum feedthrough, and down into the DAQ room where they are finally amplified before digitization.
Hotel Balmhorn, Switzerland
Fancy bivvy spot with MREs and white wine, prior to waking up at 5 AM and scrambling up the dirtiest glacier I've been on.
Gran Sasso, Italy
The XENON detector was/is located a couple thousand meters below these peaks! Great to be out on top for once.
Mönch Summit Ridge, Switzerland
Looking back over the Bernese high peaks
Cologne, Germany
My home for several years. And yes, I took this picture. Yes, I know the exact same picture has been taken by a million others.
XENONnT DAQ, Italy
A cabled-up XENONnT DAQ is nearing completion at LNGS in Italy.
Blüemlisalp, Switzerland
A nice view after a tiring climb.
Wapta Ice Fields, Canada
Purely theoretically I was attending a conference at Lake Louise at the exact moment this photo was taken.
XENON1T TPC, Italy
Don't worry. They didn't let me touch anything.
