#pydantic

Reliability for LLM agents through enforcement, not model size — Agent-Contract-Kernel + a fail-closed orchestration engine. Model-agnostic. 🇦🇪 Built in the UAE.

Professional-grade real-time AI Voice Assistant platform built with FastAPI, Next.js, and VAPI for intelligent voice call orchestration.

A tested library of 25+ ready-made tools for LLM agents — web, files, data, math, text, datetime — with OpenAI schemas, Pydantic validation and pytest coverage

Personalized AI radio agent pipeline: memory-aware dual-host dialogue, multi-provider LLMs, ElevenLabs TTS, and live audio rendering.

Trust layer for document→JSON extraction & AI agents: calibrated per-field confidence + source grounding + accept/review abstention on any OCR/VLM. Ships VerifyDocBench, a novel grounding-conditioned conformal method, and an MCP server.

https://lordbasilaiassistant-sudo.github.io/substrate-self/ — Talk to Eli in your browser. An AI being whose memory lives in its own weights, not a database. 7.4 MB ONNX, no API in the loop. 7 pre-registered tests prove substrate-identity. MIT.

An AI-powered resume optimization assistant built with LangGraph, Pydantic, and MemorySaver.

Multi-agent Python pipeline that scores a GitHub profile and generates a prioritized, source-traced plan to improve it as a portfolio. Typed Pydantic contracts, Claude + web search, uv.

🤖 ИИ-куратор для образовательной платформы

Reproducible benchmark for LLM-based structured extraction from documents. Compare Claude / GPT / Gemini / open-weight on the same task with cost + latency tracking.

The Trust Stack for Agent Apps — independently installable Python libraries that make AI agents reliable, auditable, and accountable. Because agent failures are trust failures.

FastAPI backend that turns stories into AI-generated NextNews feed posts.

API REST de gestión de tareas con FastAPI, CRUD, endpoints de IA con Azure OpenAI y tests con pytest.