Dr. Viacheslav Dubrov

Building core AI infrastructure powering autonomous agents across HubSpot's product suite. Currently on the Agent Execution team — working on LLM serving, agent evaluation, and optimizing agent runtime performance at scale. Previously led the AI Signals team — designed and shipped retrieval, grounding, and context infrastructure for AI agents, including a company-wide Embedding Hub (vector search platform on Qdrant).

Explore and create statistical and machine learning models for text preprocessing.

Devised interactive dashboards for managers based on Scala Spark and Apache Zeppelin. Researched and optimized the ML models for churn issue and text clustering.

- Engineered and implemented identification method of incidents in network traffic based on ML. - Developed and integrated Clustering model for network traffic threats database.

- Formalized and programmed methods for intelligent predictive control of electromagnets. - Implemented electrical steel control system based on statistical and machine learning models.

- Devised and integrated methods in the monitoring system of the digital substations based on ML. - Implemented algorithms and engineered NMR spectrometer. - Developed methods and diagnostic system of high-voltage breakers based on ML.

Intern Research (Lomonosov Scholarship)

Machine learning, Neural Networks, Genetic algorithms, Data Science, Diagnosis

Information measurement technique and technology, programming, statistics