Software Engineering + AI/ML | Data Science @ UC Berkeley | RISE Global Winner

• 🔭 Building TacticSquare for UAV coordination simulation and FireImpactology for wildfire exposure analysis.
• 🎯 Focus: multi-agent RL, stochastic combat models, graph optimization, remote sensing.
• 🛰 Working with Sentinel-2 and ERA5 data, downscaled climate projections and geospatial pipelines.
• 🇺🇦 From Ukraine; my projects span heat risk, wildfire exposure and UAV coordination.

• Waves (live demo) - FastAPI heatwave map for Ukraine with deterministic personal-risk scoring and Mistral-generated bilingual action plans under privacy and medical-safety guardrails.
• heat-risk-cuda - custom CUDA C++ kernels compiled at runtime with NVIDIA cuda.core, verified for exact parity with NumPy and PyTorch references (171 GPU tests) and benchmarked on a Tesla T4: a shared-memory tiled kernel cut end-to-end time 2.2× and runs about 5× faster than the PyTorch baseline.
• SwarmDAQ (live demo) - performance exchange for AI agents built at WeaveHacks 2025: agents bid on tasks, a market maker routes work to the best swarm, and reputation updates after every run, with W&B Weave tracing and Redis memory.
• LargeVCModel - Next.js + PostgreSQL platform for relationship search, company research, investor-fit scoring and secure Google integrations.
• TacticSquare - AI combat simulator for UAV coordination, route planning and attrition forecasts using RL, graph search and Bayesian Monte Carlo; field-tested in Ukraine with 10 divisions and backed by external funding.
• FireImpactology - wildfire analytics pipeline: MODIS/VIIRS/Sentinel-2 + ERA5 + Gaussian plume to map PM2.5/PM10 and health risk.
• Schmidt DSE Climate Pipeline - xarray and Dask workflows for downscaled CMIP6 analysis and National Park Service climate-risk maps at the Eric and Wendy Schmidt Center for Data Science & Environment.
• DEMI - wearable prototype with gas sensors, microcontroller and CV backend that detects TVOC spikes and automatically triggers a deodorizing spray.

15 merged pull requests in established open source projects:

• OWASP Java Encoder · added the new Encode.forJson API, escaped DEL and C1 controls in the JavaScript encoders, added XML 1.1 tags and EL functions, propagated the forUri deprecation and fixed OSGi import ranges
• GeoPandas · made the explode index dtype consistent across platforms
• All the Places · fixed missing Coles stores and added spiders for Royal Mail, Foncia, Promod, Lissac, Jean Louis David, Roc Eclerc and Monceau Fleurs covering 139,000+ locations
• OWASP CVE Lite CLI · refactored the install command

In review at OWASP DependencyCheck, GeoPandas, rioxarray and Kigali Sim.

• RISE Global Winner (Schmidt Futures & Rhodes Trust), 2024
  Top 100 of 50,000+ from 170+ countries with scholarship and project funding.

• King Abdulaziz and Mawhiba Award, 2024
  Top 10 of 1,500 finalists for a combat-operations simulator using Monte Carlo and cellular automata.

• Genius Olympiad Gold (Environmental Science), 2024
  "FireImpactology" for wildfire, air quality and health impact modeling.

• INFOMATRIX International - Platinum Medal, 1st place among 300 participants
  Advanced combat simulation with stochastic Lanchester models and RL-aided route planning.

• Venture and deployment
  $50,000 raised to scale TacticSquare pilots with 10 divisions and international visibility.

Core domains: RL (DQN, PPO), multi-agent systems, Monte Carlo, stochastic processes, graph algorithms, geospatial and remote sensing (GeoPandas, Rasterio, rioxarray, GDAL), time series, simulation.

• Resume: View PDF
• Website: vborysenko.com
• LinkedIn: Volodymyr Borysenko
• Email: volodymyr_borysenko@berkeley.edu