AI Engineering & Research
From search systems that understand meaning to models adapted for your data and research into efficient generative AI.
Send us an email →Build search that understands what people mean, not just the words they type. I help design and improve embedding-based retrieval and ranking systems, from model selection and data preparation to evaluation against real search needs.
Customer value: Help users find relevant results more easily, improve search quality, and give your team clear ways to measure progress.
Adapt pretrained models to your domain, language, and tasks. I can help assess your data and goals, select a suitable base model, build the fine-tuning pipeline, and evaluate the results.
Customer value: Get a model better suited to your use case, with improvements measured against your business goals.
Explore efficient generative models with low-rank tensor methods. My research focuses on Tensor Trains, DMRG, and rectified flow matching—approaches for representing complex data and generative dynamics with compact models.
Customer value: Investigate ways to reduce model size and computational demands, making advanced generative modeling more practical to test and apply.
BetaFlow is the independent AI engineering and research practice of Mohamed Baddar. My work connects hands-on experience in semantic search and ranking with doctoral research in low-rank tensor methods and generative modelling.
Every project starts with a clear problem, a measurable baseline, and experiments that guide the next decision.
Tell me about your problem, your data, and what success looks like.
Send us an email →