We can organize an interview with Aldin or any of our 25,000 available candidates within 48 hours. How would you like to proceed?
Schedule Interview NowMy name is Abdelilah Y. and I have over 0 years of experience in the tech industry. I specialize in the following technologies: Machine Learning, Artificial Intelligence, Data Analysis, Retrieval Augmented Generation, Computer Vision, etc.. I hold a degree in Master of Science (MS), Master of Engineering (MEng), . Some of the notable projects I've worked on include: Local Multimodal RAG, Style Transfer, Fine-tuning a small language model for summarization. I am based in Les Ulis, France. I've successfully completed 3 projects while developing at Softaims.
I am a dedicated innovator who constantly explores and integrates emerging technologies to give projects a competitive edge. I possess a forward-thinking mindset, always evaluating new tools and methodologies to optimize development workflows and enhance application capabilities. Staying ahead of the curve is my default setting.
At Softaims, I apply this innovative spirit to solve legacy system challenges and build greenfield solutions that define new industry standards. My commitment is to deliver cutting-edge solutions that are both reliable and groundbreaking.
My professional drive is fueled by a desire to automate, optimize, and create highly efficient processes. I thrive in dynamic environments where my ability to quickly master and deploy new skills directly impacts project delivery and client satisfaction.
Main technologies
0 years
1 Year
1 Year
1 Year
Potentially possible
Idemia
Built a completely local RAG solution using llama.cpp, Qwen models, and Milvus—zero API calls, 100% data privacy. Processes PDFs, text, and audio files with semantic search and intelligent Q&A. Integr
Implemented feed-forward neural style transfer based on Johnson et al. (2016) using PyTorch. Built transform network with instance normalization and residual blocks, trained on COCO dataset with perce
Fine-tuned Qwen2.5-0.5B-Instruct on CNN/DailyMail dataset (10K samples) using LoRA and Unsloth for efficient training. Implemented supervised fine-tuning with response-only loss masking. Evaluated usi
Master of Science (MS) in
2024-01-01-2025-01-01
Master of Engineering (MEng) in
2021-01-01-2024-01-01
in
2019-01-01-2021-01-01