◆ Case studies
Systems we shipped, and what they changed
Four production AI builds across healthcare and FinTech. Every figure on these pages is a measured result from delivered work, not a projection.
-
Ace Money Transfer · FinTech
KYC onboarding, fully automated
A full-stack identity-verification system pairing OCR with facial recognition to onboard customers without manual review, plus a lightweight YOLOv8n document classifier for edge deployment.
- €40k annual cost removed
- 98% field-detection accuracy
- 70% less compute vs. baseline
- PaddleOCR
- DeepFace
- YOLOv8n
- OpenCV
- FastAPI
-
OphyCare · Healthcare
One conversational API over four agents
A LangGraph ReAct orchestrator over GPT-4o that routes clinicians’ natural-language requests to specialized sub-agents for clinical notes, insurance eligibility, records RAG and routing, collapsing scattered endpoints into a single API with persistent per-user memory.
- 4 sub-agents unified
- 80% code cut in refactor
- 0 citation hallucinations
- LangGraph
- GPT-4o
- pgvector
- PostgreSQL
- AWS
-
EdgeCIO · FinTech
Graph + vector retrieval platform
A dual-module retrieval platform combining Neo4j knowledge graphs with MongoDB vector embeddings, orchestrated with LangChain and served over gRPC microservices with Redis caching.
- +40% query accuracy
- +35% answer relevance
- <200ms end-to-end latency
- Neo4j
- MongoDB
- LangChain
- gRPC
- Redis
-
OphyCare · Healthcare
AI medical scribe & form autofill
An agent that transcribes patient-doctor audio and auto-populates 10+ mandatory medical forms with per-field confidence scoring for clinician review, removing hours of documentation per provider, per day.
- 10+ forms auto-filled
- real-time speech-to-text
- Deepgram Nova-2
- LangGraph
- OpenAI
- FastAPI
Want results like these?
Tell us the problem. We will tell you honestly whether it is worth building.