Hi, I’m Aser 👋
MSc in Mathematics & Machine Learning. I work on Visual Reasoning, Reinforcement Learning, and elegant bridges between theory and application, especially with multimodal models.
“Tiny consistent steps compound into transformations that once seemed impossible.”
I’m actively seeking job opportunities and open to research collaborations. Don’t hesitate to get in touch or explore my portfolio to see more of my work!
🎯 Interests
- Applied Mathematics, AI/ML (Reasoning, Foundation Models, Reinforcement Learning), AI Fairness, AI Safety
📄 Research
- Boammani Aser Lompo, Marc Haraoui MODALITY-SWAP DISTILLATION: Rendering Textual Reasoning into Visual Supervision. (submission soon) — PDF
- Boammani Aser Lompo, Marc Haraoui Visual-TableQA: Open-Domain Benchmark for Reasoning over Table Images. NeurIPS 2025 Workshop — arXiv
- Boammani Aser Lompo, Thanh-Dung Le, Philippe Jouvet, Rita Noumeir Multi-objective Representation for Numbers in Clinical Narratives: A CamemBERT-Bio-Based Alternative to Large-Scale LLMs. CoRR 2024 — DOI
- Camille-Sovanneary Gauthier, Romaric Gaudel, Elisa Fromont, Boammani Aser Lompo Parametric Graph for Unimodal Ranking Bandit. ICML 2021 — HAL
-
Boammani Aser Lompo Impact of Spurious Correlation on Gradient Descent Dynamics and Generalization. (submission soon) — GitHub
- Reviewer: ICCV 2023 • CVPR 2026 • ICLR 2026
🧩 Featured Projects
🛡️ Youth Mental Health Safety Guardrail
Fine-tuned Qwen2.5-7B + LoRA · 9-signal distress taxonomy · F1=0.85
🧾 Visual-TableQA Pipeline
Multimodal reasoning · 2 500 LaTeX tables · 9 000 QA pairs · LLM-jury QC
🧬 Multi-Objective Token Representation
LESA + Xval · Numerical reasoning · CamemBERT-Bio
⚖️ Optimal Transport for Color Transfer
Unbalanced OT · Color distribution transfer · Visualizations
♟️ Reinforcement Learning for Chess
C++ · SARSA + Monte Carlo · ~1350 Elo after 1h self-play
🧠 Tech Stack
- Languages: Python · Model Context Protocol (MCP) · JavaScript · C++ · Bash · HTML/CSS
- Frameworks: PyTorch · Transformers · JAX · Hugging Face · APIs
- Tools: Docker · Paperspace
- Focus Areas: Visual-Reasoning · Number Encoding · Reinforcement Learning · Applied Maths
Latest Posts
When an AI assistant talks to young people in distress, the stakes are high. A response that misses a cry for help — or that over-triggers on benign conversa...
In this short tutorial, we’ll walk through how to fine-tune a Vision–Language Model (VLM) on a custom dataset. Our case study will use Qwen2.5-VL-7B-Instruct...