Abdeslem Hmn

I've built AI and optimization systems spanning energy-aware MaxSAT solving, hardware-constrained neural architecture search, and medical-data ETL pipelines.

For GreenACO-MaxSAT, I designed an eco-energy-aware ant colony optimization solver with UCB1-controlled search regimes, hardware energy instrumentation through pyRAPL, CO₂ tracking with CodeCarbon, and execution tiers that achieved approximately 100× speedup on SATLIB and MaxSAT Evaluation 2024 benchmarks.

I also developed HW-NAS-Memetic, a hybrid memetic algorithm for finding CNN architectures across accuracy and hardware-latency trade-offs, and built a Python pipeline that collects, cleans, normalizes, and consolidates biomedical disease data for querying and machine learning.

Alongside my engineering work, I've developed Django backends at Ourquilane, delivered complete freelance web solutions, and contributed to the AI community as a DataHack 3.0 challenge author and mentor and Vice-President of the ESI Scientific Club.