About Shaheryaar
I help startups and small companies turn ideas and data into working AI solutions.
I am an AI Engineer with hands-on experience in Machine Learning, Computer Vision, and Python-based backend systems. I specialize in building practical, production-ready ML features from data preparation and model training to API integration and deployment.
My recent work includes computer vision pipelines (YOLO, CNNs) for defect detection, ML-based classification systems, and end-to-end prototypes deployed with FastAPI and Docker on cloud platforms (AWS, Azure). I focus on clean architecture, reproducible experiments, and measurable results.
Typical projects I support:
• ML or computer vision prototypes (POC → MVP)
• Model training, evaluation, and optimization
• Backend APIs for AI features (FastAPI/Django)
• Data analysis and model performance reviews
I enjoy working with early-stage teams that need fast, reliable technical execution and clear communication. Happy to start with a small, well-defined project and scale from there.
English
Native or bilingual
German
Conversational
Can work on-site
München (up to 50km)
Experience
- TH RosenheimResearch Assistant – Prediction Algorithm in MobilityOctober 2025 - Today (8 months)Rosenheim, BY, Germany• • Developing ML-based mobility prediction algorithms for real-time transportation scenarios.• • Conducting data analysis on traffic and mobility datasets to improve prediction accuracy.• • Collaborating on research publications and implementing model evaluation workflows.
- AI4GreenResearch AssistantFebruary 2025 - Today (1 year and 4 months)• • Developed YOLO-based defect detection pipeline for sustainable manufacturing• • Prepared & annotated 8K+ image dataset, performed augmentation and labeling• • Fine-tuned YOLO models, improving detection accuracy (mAP50 0.85, Precision 0.85, Recall 0.78)• • Integrated deployment pipelines and ran GPU training loops
- HigHRoQ Teaching Innovation (CNN Methods)Student AssistantAugust 2024 - March 2025 (7 months)• • Built and optimized a custom CNN for image classification with efficient architecture and tuned hyperparameters.• • Applied regularization, dropout, and data augmentation to improve model generalization.• • Created clear documentation and a reusable training/evaluation pipeline.• • Designed student exercises to teach CNN training, dataset bias, and deep learning workflows.• • Integrated explainability tools (Grad-CAM, confusion matrix, metrics visualization) for model interpretation.• • Delivered a lecture on CNNs for Wood Technology students, linking AI methods to defect detection.
Recommendations
Be the first to recommend Shaheryaar
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- B.Sc. Applied Artificial IntelligenceTH Rosenheim2025B.Sc.