From a task —
to a working system.
We design and ship AI and web products to production. We take the whole product: interface, backend, AI integrations, launch and the next iteration. And we put our name on what we build.
The whole product
Product, UX, backend, AI, launch and iteration in one senior team. No seams between three contractors.
A short path to production
Not specs for the sake of specs. A minimal release with a measurable signal; architecture built to grow, not to be rewritten.
AI in production, not in a demo
LLM features, assistants, RAG — with safety and cost control, for a specific measurable outcome.
The workshop's expertise
AI interfaces & assistants
LLM features, personalized flows, RAG, event analytics, safe AI integration into a real product.
Landing & product sites
Marketing and product pages that explain value and bring leads.
Internal panels & platforms
CRM, admin panels, dashboards, knowledge bases — operational interfaces for daily work.
Integrations & automation
Processes across services, payments, analytics, CRM and internal systems — with no manual gaps.
Technical launch
CI/CD, Docker, production config, monitoring, infrastructure — a product you can run and maintain.
Architecture for growth
Project, content and data structured so the product can grow.
Non-trivial domains, shipped
NovaMed
AI-based clinical decision support: multimodal analysis of history, labs and records; probabilistic recommendations — in a sensitive domain with safety.
EyeMesearing
Automated calculation of ophthalmic parameters from an eye photo: MRD1, MRD2 and others, segmentation, integration with clinics and telemedicine.
Pollution Detection
Oil-spill detection from satellite and UAV data: online detection, map output, analytics. Result: response delay from 6–48 h to online.
We hone our craft on our own products
Alongside client work we build our own products — it is part of the model, not a side gig. It keeps us in real product discipline: architecture, analytics, growth, safety, the production cycle.
An AI diary for self-reflection
Voice and text entries, structuring questions, dynamics over time. Closed beta, a separate brand. For ИОАНН it is proof of the ability to take 0→1 to production.
A separate B2C brand · inner-voice.tech