https://github.com/maxeq/maxeq
Profile README
https://github.com/maxeq/maxeq
Last synced: 1 day ago
JSON representation
Profile README
- Host: GitHub
- URL: https://github.com/maxeq/maxeq
- Owner: maxeq
- Created: 2026-06-19T13:08:43.000Z (10 days ago)
- Default Branch: master
- Last Pushed: 2026-06-19T14:02:37.000Z (10 days ago)
- Last Synced: 2026-06-19T15:30:28.487Z (10 days ago)
- Size: 1.95 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
### Hi, I'm Maksims 👋 — Applied AI / Identity-Security Engineer
I build the full trust pipeline for **biometric identity** — from the ECG signal and the ML model that reads it, through cryptographic keys that never touch disk, to the OIDC/SAML tokens that prove it. Most teams need three specialists for that seam; I've shipped all three in production.
🔭 Core engineer at an ECG-biometric authentication startup — identity, ML, and desktop.
🌍 EU citizen · remote across the EU · open to identity-security / applied-AI / full-stack roles.
**🔐 Identity & security** — OIDC/SAML identity provider (node-oidc-provider), biometric proof-of-possession (DPoP, RFC 9449), 3-shard secp256k1 key derivation, Microsoft Entra federation, Azure (Key Vault, Container Apps).
**🧠 Applied & biometric ML** — metric learning (ArcFace / SubCenter ArcFace / SupCon) on InceptionTime, ONNX with score calibration baked into the graph, on-device inference, Modal serverless GPU.
**⚡ Real-time / AI products** — OpenAI Realtime voice, RAG over pgvector, WebSockets/SSE. Built **Diligentify** — a live AI voice-interview SaaS, solo end-to-end.
### Selected projects
| Project | What it is |
|---|---|
| [diligentify-case-study](https://github.com/maxeq/diligentify-case-study) | Live AI voice-interview SaaS — OpenAI Realtime, RAG/pgvector, Stripe |
| [polyphase-resample](https://github.com/maxeq/polyphase-resample) | Bit-exact cross-platform DSP resampler (matches scipy to float64) |
| [dpop-proof-of-possession](https://github.com/maxeq/dpop-proof-of-possession) | Sender-constrained tokens — DPoP (RFC 9449) |
| [onnx-confidence-calibration](https://github.com/maxeq/onnx-confidence-calibration) | T-norm + confidence calibration baked into an ONNX graph |
| [datatable](https://github.com/maxeq/datatable) | Published React data-table — npm `@maxeq-ui/datatable` |
**Stack:** TypeScript · Python · Node.js · Next.js · ONNX / TensorFlow · secp256k1 · OIDC/SAML · Azure · Docker
🔗 [diligentify.com](https://diligentify.com) · [Portfolio](https://portfolio-maxeqs-projects.vercel.app) · [LinkedIn](https://www.linkedin.com/in/maksims-volo%C5%A1%C4%8Duks-13152656/)