Bogdan and team.

Real data complexity, live state, accumulated history; operations that have to make decisions on what they have. Customer intelligence, social-media intelligence, field operations, premium consumer apps with real backend depth.

Bogdan is the principal — most of the architecture and most of the code that ships from here ships through his keyboard. The motivation came out of years on the wrong side of two recurring problems: shops that shipped great-on-a-laptop systems that fell over the day a customer reported missing data, and clients who wanted a "simple MVP" for what was actually a full business. The first taught the production-seriousness lane. The second taught the for operations that already exist filter.

What holds our attention is the layer between data and decisions: orchestration that survives partial failures, ML that ships and stays shipped, intelligence layers that turn existing operations into ones that can answer questions of themselves. Async-first, working with teams who have stakes.

What we ship on. Python and C++ are both first-class. Django + DRF on the backend; Next.js for product UIs and content surfaces; SwiftUI for native iOS. Data: PostgreSQL, Redis, Celery, Django Channels. ML: PyTorch, TensorFlow, OpenAI and Anthropic APIs. Deployed to Hetzner, AWS, Vercel. The list is representative, not the whole toolkit.

Adjacent: Bogdan publishes the Bogdan's Lab newsletter and writes longer-form on bgdnandrew.com. Other projects — QuantSandbox, MLSandbox, SystemSim, and the bogdan.engineer hub — are linked in the footer.