+971566341484 | vladserkoff@pm.me | linkedin.com/in/vladserkoff | github.com/vladserkoff
Senior AI/ML Engineer specializing in computer vision, NLP, and artificial intelligence applications. Proven track record deploying production ML systems, optimizing model performance, and building efficient data pipelines. Based in Abu Dhabi, UAE, open to remote work and relocation.
- Established MLOps workflows and development processes for the team
- Fine-tuned LLMs for cybersecurity domain, automating threat detection workflows
- Built typosquatting detection system to protect client domains from malicious registrations
- Co-developed graph neural network architecture that reduced 3D point cloud processing time by 50%
- Redesigned training infrastructure, cutting model development cycles from 7 days to 20 hours and reducing AWS costs by 80%
- Created hyperspectral tree species classifier achieving 30% higher accuracy than RGB baseline
- Implemented LiDAR-RGB fusion pipeline, reducing object detection false negatives by 30%
- Streamlined annotation workflows, increasing labeling team efficiency by 50-70%
- Designed time-series forecasting neural network surpassing human accuracy benchmarks
- Mentored intern through successful transition to full data scientist role
Led ML initiatives across Stream, DCA, Flocktory, and Mamsy, focusing on real-time bidding optimization, NLP-based product taxonomy, and reinforcement learning for ad placement systems.
- Deep Learning & ML: PyTorch, Scikit-learn, HuggingFace Transformers, AWS SageMaker
- Computer Vision & NLP: Object detection, segmentation, sensor fusion, OpenCV, BERT, LLM fine-tuning
- Infrastructure & Tools: Python, SQL, Docker, Git, PostgreSQL, vector databases, CI/CD
MSc Economics, HSE (2013) | Co-author: ECLAIR LiDAR Dataset, CVPR 2024 Workshop