Atik Islam Sajib
I am a Machine Learning Engineer with four years of experience designing, building, and deploying ML systems.
About
I am a Machine Learning Engineer with four years of experience designing, building, and deploying ML systems. I specialize in computer vision, multimodal learning, and large language models, and I enjoy turning complex data into practical, scalable solutions. My current work focuses on satellite radar data fusion for precipitation nowcasting at Fraunhofer HHI, where I drive end-to-end ML experiments and production-ready pipelines. I’m passionate about robust architectures and cross-disciplinary collaboration, and I love sharing knowledge to accelerate impactful results.
Beyond research, I’ve built end-to-end data pipelines in Python and SQL, mentored data annotation workflows, and contributed to open ML tooling. I hold an MSc in Data Science (expected 2026) from Berliner Hochschule für Technik and a BSc in Computer Science & Engineering from United International University, Dhaka. I’m fluent in English and Bengali, with working German and Hindi skills, and I’m excited to apply ML to real-world problems.
- 6 years' experience
- Joined March 2026
Languages
- EnglishFluent
- BengaliFluent
- HindiIntermediate
- GermanAdvanced
Industry experience
- Software & Internet
- Media & Entertainment
- Education
- Professional Services
Work experience
Architected automated pipelines and contributed to real-time high-precision nowcasting using RADOLAN/EUMETSAT data. Led multimodal ML experiments and implemented scalable pipelines with Git, DVC, Docker, MLflow, and REST APIs to support high availability and reproducibility. Coordinated with project stakeholders to align model performance with KPIs and research objectives.
Developed computer vision systems for vehicle classification with real-time monitoring (94% accuracy). Built NLP models with Transformer-based approaches and fine-tuned large language models for Bengali text analysis. Designed scalable Python–SQL data pipelines for large ML datasets, and managed annotation workflows (LabelMe/LabelImg). Implemented Bangla text preprocessing (NLTK/BNLP) improving sentiment classification by ~12%.
Education
Hire Atik Islam Sajib today
Post a job and invite Atik, or contact them directly. No hidden commission.