Team / profile
Mislav Jordanić
Data scientist · PhD, Biomedical Engineering · Zagreb, Croatia
Profile
I am a data scientist with a background in biomedical engineering and more than ten years of work with data and models. Most of my recent work has been demand forecasting, prediction, and inventory and price optimisation for clients, including combining LLM-based agents with classical models.
Before moving to industry I spent seven years as a researcher and lecturer at the Universitat Politècnica de Catalunya, working with biomedical signals, mainly high-density EMG: recording muscle activity and using it to identify movements and estimate force, for rehabilitation and human-machine interfaces.
Since 2025 I have been working at Lamis Technologies.
Experience
- 2025–present
Data scientist
Lamis Technologies · Zagreb, Croatia
Data science and machine learning work for clients: forecasting, prediction and optimisation.
- 2022–2025
Machine Learning Scientist
Pendulum Systems · Washington, DC, USA
Demand forecasting and inventory optimisation for clients in several regions.
- Built forecasting models for client data that was often sparse or noisy.
- Set up model evaluation, tracking and explainability reporting so planners could see why a forecast moved.
- Worked on price and inventory planning with constrained optimisation.
- Worked on simulation for supply chain and inventory planning.
- Contributed to an internal Python library for demand forecasting.
- Combined LLM agents with classical forecasting models, and developed ways to evaluate the stability of the combined system.
- Built an LLM-based data ingestion flow that reduced manual data preparation.
- Worked on delivery-time and location estimation from traffic and route data.
- 2021–2022
Data Scientist / Quantitative Researcher
Aprawa · Cheyenne, WY, USA
Trading strategies built on gradient-boosted trees (CatBoost, XGBoost) and sequence models (Keras).
- 2014–2021
Researcher and Lecturer, Biomedical Engineering
Universitat Politècnica de Catalunya (UPC) · Barcelona, Spain
Acquisition and analysis of biomedical signals (EMG, ECG, EEG, EOG, IMU, force), with a focus on predicting limb force and motor tasks from muscle activity.
- Took part in five research projects and led the technical side of three innovation projects, including a 256-channel wireless EMG amplifier.
- Took a technology-transfer project through international healthcare accelerator programmes.
Skills
Modelling
- Forecasting and time series
- Gradient-boosted trees (LightGBM, XGBoost, CatBoost)
- Deep learning (PyTorch, Keras)
- Reinforcement learning (PPO)
- Statistical methods
Optimisation and simulation
- Constrained optimisation (cvxpy, MOSEK)
- Hyperparameter search (Optuna)
- Discrete-event and agent simulation
- Inventory policies
LLMs and agents
- LangGraph
- Retrieval-augmented generation
- OpenAI and Anthropic APIs
- Evaluating LLM-assisted systems
Signals and data
- Digital signal processing
- EMG, IMU and other biosignals
- SQL
- Geospatial data (GeoPandas, OpenStreetMap, Valhalla)
Engineering
- Python, C, MATLAB
- pytest, git
- MLflow, ClearML
- AWS (SageMaker, Lambda, EC2, S3, ECS), Google Cloud Run
- Docker
Education
- 2019
Postgraduate specialisation in Data Science and Big Data
Universitat de Barcelona · Barcelona, Spain
- 2017
PhD in Biomedical Engineering
Universitat Politècnica de Catalunya · Barcelona, Spain
Thesis on muscle activity patterns from multichannel surface EMG; supervised by Miguel Ángel Mañanas and Mónica Rojas-Martínez.
- 2012
MSc in Electronic and Computer Engineering
University of Zagreb, Faculty of Electrical Engineering and Computing · Zagreb, Croatia
Publications
Research from my time in biomedical engineering, mostly on high-density EMG. Profiles: ORCID, OpenAlex.
Journal articles (7)
- 2022
- 2020
- 2019
- 2017
- 2017
- 2016
- 2016