I’m Aml Hassan, a Data Scientist passionate about building AI systems that make a real impact. My journey into AI started in 2021, and since then I’ve worked across machine learning, NLP, TTS, computer vision, and federated learning, tackling challenging data problems and experimenting with cutting-edge models.
- 💼 Currently working as an Applied Scientist at Microsoft, improving product experiences by fine-tuning multimodal models (like BLIP) and integrating AI into Copilot.
- 🧠 Previously at WideBot, I built scalable pipelines for Arabic TTS and fine-tuned models for summarization in both MSA and Egyptian dialects.
- 🌍 I’ve also contributed to open-source federated learning research with Flower Labs, and developed projects like AI Lip Sync and bilingual text clustering.
- 🔬 My interests span multimodal AI, generative models, and efficient ML systems.
- 🌱 Always learning.
- ⚡ Fun fact: Aml means Hope 😄
- Python, Git, Docker, Poetry, Hydra-config, Weights & Biases (W&B)
- PyTorch, XGBoost, SentenceTransformers, nltk, networkx
- Contrastive Learning, Few-shot Learning, Federated Learning
- NLP (Summarization, Semantic Search, Clustering)
- Computer Vision (BLIP, CLIP, ResNet)
- Generative AI (TTS, Multimodal AI)
- Text-to-Speech (XTTS V2, Fish-Speech)
- Whisper, Pydub, Forced Alignment
- Arabic NLP (MSA & Egyptian Dialect)
- Flask, Streamlit
- HTML, CSS, JavaScript
- Data Collection & Cleaning
- Model Fine-tuning & Evaluation (Cosine Similarity, AUC, Accuracy)
- Clustering (Louvain)
Check out my Medium where I share insights from my projects.
aml.hassan.esmil@gmail.com
https://www.linkedin.com/in/aml-hassan-esmail