Hamid Badri
Portrait of Dr. Hamid Badri

Dr.Hamid Badri

Lead Data Scientist, Battery Modeling & Diagnosis

Circunomics GmbH · Frankfurt am Main, Germany

I build models that tell how a lithium-ion battery is aging and how much life it has left. At Circunomics I lead the battery modeling and diagnosis work behind a platform that assesses EV and stationary batteries for second-life use and recycling.

My background is physics and numerical modeling. My PhD at the University of Strasbourg was on inverse problems: recovering hidden parameters of flow in porous media from sparse measurements with adjoint-state methods. Estimating a battery's state of health from field data is the same kind of problem, and I work on it with a mix of physics-based and machine-learning models, from data pipelines to production.

CV (PDF) LinkedIn Google Scholar ORCID ResearchGate GitHub
100% 80% 50% 1st life · EV 2nd life · storage recycling lower stress high stress low stress SoH time · cycles →
Battery aging, sketched. Gentler charging, shallower cycles and keeping the pack at a moderate temperature, neither hot nor cold, push the knee later. A pack retired from a car at about 80% can then work for years as stationary storage, until it drops below about 50% and is recycled. Second life has its own knee, so the fade speeds up again near the end.

Experience

Aug 2021 – now

Circunomics GmbH · battery lifecycle analytics & marketplace

  • Lead Data Scientist, Battery Modeling & Diagnosis · since Sep 2026Current
  • Senior Data Scientist · Aug 2021 – Aug 2026
  • Lead battery state-of-health, degradation and remaining-life modeling for EV and stationary storage batteries, in first and second life.
  • Multivariate time-series forecasting, predictive maintenance and anomaly detection with machine learning and deep learning.
  • Built and lead the data pipelines in Apache Airflow on Kubernetes (Azure AKS), and the analytics app that serves KPIs, time series and detected anomalies to customers.
  • Present the analytics to customers and drive new data-driven product features.
Jan – Jul 2021

Data Scientist · BIAL-R / Merck Group (M-Lab)

  • Multivariate time-series forecasting and predictive maintenance with deep learning and machine learning.
  • Wrapped models in REST APIs and prepared them for serving on AWS Lambda.
Mar 2019 – Aug 2020

Data Scientist · Stockbooking, France

  • Stock optimisation with reinforcement learning (Q-learning, SARSA, deep Q-networks).
  • Demand forecasting with LSTM and SARIMAX; data pipelines; API serving forecasting and decision models.
  • Managed the company's part of ICONET, an EU research project on the Physical Internet.
2015 – 2018

Doctoral Researcher · University of Strasbourg

Inverse problems and model calibration: identification of parameters and source terms for flow and transport in porous and fractured media, with continuous and discrete adjoint states and constrained optimisation. Teaching assistant at the Faculty of Physics & Engineering, 2016–2018.

Earlier, 2010 – 2015
2015 · 5 mo

Li-ion battery modeling · ICube / QEERI, Qatar

Nonlinear thermal and multi-physics model of lithium-ion batteries.

2014 · 1 mo

Consultant · Qatar Environment and Energy Research Institute

2013

Technical Manager · GREEN-CAR, Morocco

Design and build of a solar-car prototype.

2012 · 4 mo

Graduation project · OCP SA, Morocco

Raised plant availability: flocculation and decanter process, automation system and operator GUI.

2010 – 2011

Technical Manager, team of 10 · ENSAcar, Shell Eco-marathon Europe 2011

Built a four-wheel urban-concept car for the competition in Germany.

Research Inverse problems · numerical modeling

41Citations
2h-index
2018–Reviewer, Hydrogeology Journal
  1. 2019 Heterogeneous hydraulic conductivity and porosity fields reconstruction through steady-state flow and transient solute transport data using the continuous adjoint state F. Delay, H. Badri, P. Ackerer · Advances in Water Resources 8 cit.
  2. 2018 Identification des paramètres de l'écoulement et du transport en milieu poreux par la méthode de l'état adjoint continu H. Badri · PhD thesis, University of Strasbourg 2 cit.
  3. 2017 A comparison of discrete versus continuous adjoint states to invert groundwater flow in heterogeneous dual porosity systems F. Delay, H. Badri, M. Fahs, P. Ackerer · Advances in Water Resources 31 cit.
  4. 2015 Self-consistent approach to solving the 1D Thomas-Fermi equation using an exponential basis set H. Badri, F. H. Alharbi, R. Jovanovic · AIP Conference Proceedings 2 cit.

Citation counts from Google Scholar, updated .

Robotics After hours · Dr.Dsc-Dev

Outside work I am building a robot arm from scratch: designing the parts in 3D CAD, 3D printing them, and writing the control software. I document every step, from mechanics and electronics to code, and share it along with science, data science and engineering content.

The source and printable parts are on GitHub.

Flexible gripper wrapping around a mug Flexible gripper

Most watched

330K views on TikTok

Flexible gripper: a robot end effector

A soft gripper for the arm, designed in FreeCAD and printed in flexible TPU. The fingers bend around an object instead of squeezing it, so the same gripper can hold a mug or an orange.

Watch on TikTok

More from the build

Longer videos

Skills

Battery analytics

State of health and charge estimation, degradation and remaining-useful-life modeling, knee detection, rainflow cycle counting, physics-based models (PyBaMM), second-life assessment.

Machine learning

Multivariate time-series forecasting, anomaly detection, predictive maintenance, reinforcement learning, hyperparameter tuning. TensorFlow, Keras, scikit-learn.

Data & MLOps

Python, pandas, NumPy, Dask, SQL. Apache Airflow, Docker, Kubernetes (AKS), Azure, AWS, Git. REST APIs with Flask and FastAPI.

Modeling & maths

Inverse problems, adjoint-state methods, constrained optimisation, PDE-based numerical simulation, MATLAB, SciPy. Dash and React for analytics front ends.

Education

2018

PhD, Physics and Numerical Modeling · University of Strasbourg

Thesis on parameter identification for flow and transport in porous media with the continuous adjoint-state method.

2015

MSc, Computational Engineering · University of Strasbourg, ICube

2012

Engineering degree, Process, Energy & Environment · ENSA Agadir, Morocco

2009

BSc, Science & Technology (Maths, CS, Physics) · Energy · FSTS, Morocco

Languages

Frenchnative

Arabicnative

Englishfluent, professional