We're documenting how we built a production-grade realtime AI tutor
combining conversational AI, streaming voice, avatar animation, GPU
inference, and model orchestration.
This engineering case study will cover the architecture, latency
challenges, deployment strategy, and lessons learned from building a
realtime AI learning experience.
Healthcare / EHR · Case Study
The AI Layer Above the EHR
An analysis of how AI front-office products are integrating with
fragmented ambulatory EHR ecosystems — and the partner economics
emerging around that integration layer.
A decade of production engineering in regulated financial services,
carried into AI systems: regulated data, financial systems, AI
infrastructure, production reliability, audit/compliance, and
engineering leadership.
Klub — Revenue-Based Financing Platform
Scaled engineering from 2 to 70+ and led the platform from concept
to over $1B processed, $200M disbursed to 800+ brands. Built
real-time collections automation and under-24-hour credit approval
workflows. Led ISO 27001 certification and ADGM/FSRA regulatory
audit. Cut infra costs 50%.
InCred — Lending Stack
Founding engineering team. Built the lending stack now supporting
a $1.2B loan book, launched in under 4 months. Scaled engineering
from 2 to 55. Configurable product architecture, multi-bureau
credit integrations (CIBIL, Experian, CRIF, Equifax).
Lucio AI — AI Infrastructure Under Audit Constraints
Multi-agent orchestration, RAG pipeline on Turbopuffer, LLM
evaluation framework, and prompt versioning infrastructure — built
to the same audit-grade standard as the regulated financial
systems above.