Curriculum Vitae
Dmytro Petrovskyi · ML/AI Engineer · Kyiv, Ukraine
ML/AI engineer with 10+ years of experience delivering production systems across LLMs, computer vision, NLP, forecasting, and ML platform engineering. Core pattern: research-to-production delivery of systems that actually ship — not prototypes.
Skills & Domains
ML/AI LLMs, Computer Vision, NLP, Forecasting, Entity Resolution, Information Extraction
Infrastructure AWS, Kubernetes, Docker, MLOps, CI/CD, Microservices
Languages Python, SQL, Java
Data E-commerce, Social/Public Data, OSINT, Financial, LiDAR/3D
Experience
NielsenIQ
Senior ML/AI Engineer (Tech Lead)- Building multi-agent code generation system for internal DSL — LLM-as-a-Judge, agent collaboration, dynamic context auto-compaction, custom orchestration
- Compressed 300+ classification models into one via massive multitask learning — ~$100K/year cloud cost savings
- Built LLM-based extraction system for large-scale e-commerce data — R&D-grade, fully shipped to production
Defense Tech (NDA)
ML/AI Engineer → Advisor/Consultant- Face/image retrieval system across ~3B public photos — petabyte-scale visual search
- AI profiling pipeline: photo/name → compiled intelligence dossier across social platforms
- Cross-platform identity resolution — merging accounts across social networks into real-world entities
- Detection models for military objects in imagery (personnel, vehicles, equipment)
- Large-scale text classification for detecting hostile/extremist social content
Rakuten / Rakuten Intelligence
Senior ML Engineer / Team Lead- Forecasting systems for 10M+ consumer panel
- Automatic information extraction for e-commerce using deep learning/NLP
- ML platform: serving, training, and monitoring on AWS/Kubernetes/microservices
- Team leadership: goal setting, coordination, business alignment
Whirl Software
ML Engineer / Team Lead- Social media monitoring system for large-scale public data
- NLP models: sentiment analysis, semantic similarity, named entity recognition
- Integrated ML/deep learning solutions into big-data application architecture
Infopulse
ML Engineer- Research and PoC for LiDAR / 3D point-cloud semantic segmentation
Zoral
ML Engineer- Financial-risk ML: credit scoring, fraud detection, bank transaction categorization
- NLP R&D: sentiment analysis engine, NER, coreference resolution
Promptlink Communications
Software Engineer- Cable network monitoring system for modems
- Statistical/early-ML approaches for network analytics and fault analysis
- Low-level networking algorithm development and optimization