Mohsen Pirmoradiyan

AI Lead & Computer Vision Engineer

Arboair AB, Linköping, Sweden · drone and aerial analytics for forestry

• Designed and lead the company's computer-vision platform: one config-driven system covering data ingestion, training-data preparation, model training, geospatial evaluation and GPU inference in production.
• Train modern deep-learning segmentation models for tree species and forest-damage mapping, from transformer-based architectures with foundation-model backbones to classic CNNs, across semantic, instance and panoptic tasks.
• Built self-supervised pretraining on unlabelled aerial imagery and semi-supervised pseudo-labelling to make the most of scarce annotations.
• Built the training-data foundation: a versioned data lake with provenance, intake of finished annotation jobs, reproducible dataset recipes and leakage-safe splits.
• Turn pixel predictions into GIS-ready maps: tiled inference over very large GeoTIFFs, vectorisation, geometry repair and certainty scoring.
• Contributed to the orthomosaic pipeline: LiDAR terrain models, direct georeferencing from drone metadata, seamline mosaicking and photogrammetric true orthophotos.
• Led the tree stem-diameter estimation project end to end, from field reference data and aerial height data to a production model with staged releases.
• As AI lead, set up an AI-agent-assisted engineering workflow and designed an AI-agent research process supporting the company's strategy work.

Master's Thesis, Ericsson Research

Ericsson AI Research, with Linköping University

• Built a random-forest classifier that flags moments of high end-to-end latency in measurements from Ericsson's 5G test network.
• Used causal Bayesian networks (structure learning with the PC algorithm, then causal-effect estimation) to find which network measurements actually drive the latency, not just move with it.
• Thesis: Anomaly detection and causal inference in 5G mobile networks, Linköping University, 2021. Summary and report.

Seismic Data Acquisition QC Supervisor (junior to senior)

National Iranian South Oil Company (NISOC), Khuzestan, Iran

• Quality control of 2D and 3D land seismic acquisition in deserts, sand dunes and mountains: recording quality, geometry and field procedures.
• Supervised acquisition in more than five oil fields, with dynamite and vibroseis sources; reviewed survey design and field data before processing.
• Mentored junior geophysicists; recognised as one of the company's best seismic acquisition supervisors in 2015.

MSc Statistics and Machine Learning

Linköping University, Sweden

Two-year programme (120 ECTS); thesis with Ericsson Research (see Experience).

MSc Geophysics (Seismology)

Institute of Geophysics, University of Tehran, Iran

Thesis: Petroleum reservoir characterization using seismic data analysis (prestack seismic inversion).

BSc Petroleum Exploration Engineering

Petroleum University of Technology, Ahvaz, Iran