About Me
👋 Hi, I'm Slava Dubrov — a Machine Learning Tech Lead with 10+ years of shipping production AI systems.
Why You Should Read This Blog
I write about what actually works in production AI systems — not theory, not hype, but patterns tested at scale. My perspective comes from:
- Building the Context Layer for AI Agents at HubSpot — designing the infrastructure that keeps agents grounded, efficient, and compliant
- Saving companies millions — $4M+ yearly at Wayfair alone through fraud/scam detection and embedding systems I built and led
- Publishing research — PhD in AI diagnostics, peer-reviewed publications, and patents in intelligent systems
- Hands-on with the full stack — from data pipelines and model training to evaluation, deployment, and safety guardrails
- Cloud ML architecture expertise — designing and deploying AI systems across AWS and GCP as batch, streaming, and real-time services
When I share a technique like Schema-Guided Reasoning or Context Engineering patterns, it's because I've validated it in real systems with real stakes.
Current Focus
At HubSpot (Tech Lead ML II), I'm building:
- Embedding Hub — production vector infrastructure on Qdrant powering semantic search and personalization
- Context Layer for AI Agents — building systems that generate actionable context and insights for HubSpot's agents using a combination of retrieval, embeddings, and reasoning techniques
I lead cross-functional ML teams, mentor engineers, and partner with product and go-to-market stakeholders to launch customer-facing AI features globally.
Career Highlights
| Company | Role | Impact |
|---|---|---|
| HubSpot | Tech Lead ML II | Built Embedding Hub and Context Layer for AI Agents |
| Wayfair | Senior ML Scientist | Led fraud ML and behavioral embeddings systems generating $4M+ annual savings |
| OLX Group | Data Scientist | Developed series of recommender systems (item2vec, ALS, deep learning, matrix factorization) |
| Zalando | Research Engineer | Demand forecasting, Spark ML pipelines |
Education: PhD in AI/Control Systems (SRSPU/TU Ilmenau) · Nanodegrees in Computer Vision, Deep RL, AI Trading
Recognition: Leonhard-Euler Scholar, UMNIK Scholar, published patents and research papers
What I Write About
My posts focus on practical AI engineering:
- Context & Retrieval — Context Engineering for agentic systems, RAG patterns, memory management
- LLM Development — Fine-tuning guide, Schema-Guided Reasoning, model serving
- Agent Architecture — Domain-Driven Design for AI Agents, orchestration patterns
- Developer Tooling — Python setup, uv, MCP servers
Tech Radar
Cloud & Infrastructure: AWS (Certified) · GCP/Vertex AI · Batch & Streaming Pipelines · Real-time ML Serving
Core: Python · PyTorch · JAX · Spark · FastAPI · Airflow · DBT
Vector & Retrieval: Qdrant · Faiss · Embeddings · Hybrid Search · Reranking
LLM & Agents: LangChain · LangGraph · LlamaIndex · CrewAI · Google ADK · FastMCP · SmolAgents
Serving & Fine-tuning: vLLM · LoRAX · Unsloth · Axolotl
Let's Connect
Let's connect — especially if you're tackling hard ML problems at scale.