You will get OmniIntelOS — Full-Stack AI Company Intelligence Platform (Palantir-Style)
Project details
A complete, self-hostable AI company intelligence platform — 9 executive AI personas, real-time data pipelines, OCR, voice, n8n automation, and a live monitoring stack. Built to run on your infrastructure.
Machine Learning Tools
BERT, GitHub Copilot, Microsoft Excel, OpenCV, Python, Python Scikit-Learn, PyTorch, scikit-learn, TensorFlow, Word2vecWhat's included
| Service Tiers |
Starter
$1,500
|
Standard
$3,500
|
Advanced
$6,500
|
|---|---|---|---|
| Delivery Time | 10 days | 21 days | 35 days |
Number of Revisions | 2 | 3 | 5 |
Model Validation/Testing | - | - | - |
Model Documentation | - | - | - |
Data Source Connectivity | - | - | - |
Source Code | - | - | - |
Optional add-ons
You can add these on the next page.
Extra AI persona
(+ 2 Days)
+$250
Additional n8n workflow
(+ 2 Days)
+$200
Technical documentation package
(+ 3 Days)
+$350Frequently asked questions
About Yacine
AI Systems Engineer |RAG | MCP Agents | Document AI | LLMOps | FastAPI
Niamey (Commune II), Niger - 12:32 am local time
Over the past year I've shipped six tools, each one solving a real problem I kept seeing across client projects. Here's what's live:
IntelAI — a persona-aware AI analytics backend with 9 C-suite AI roles across 7 business domains (Finance, HR, IT, Ops, Logistics, ESG, Risk). Each persona has its own data scope, system prompt, and tool whitelist — so the CFO and CHRO get genuinely different answers from the same system. GraphRAG-lite + hybrid retrieval + BGE reranker. vector RAG (Chroma dev / Qdrant prod) with source citations, WebSocket streaming, ML forecasting with Monte Carlo confidence bands, anomaly detection, board-ready PDF export, full RBAC, bilingual EN/FR React frontend. Self-contained cloud deployment — single app, no platform dependency. Live demo available.
AgentKit— an open-source MCP server that exposes business intelligence to Claude Desktop, Cursor, or any LangChain agent. Six tools, a three-agent LangGraph workflow for autonomous analysis, and demos for Claude Agent SDK, CrewAI, and DSPy. One JSON config and your AI assistant can query live PostgreSQL data.
DocIntel — a document AI pipeline with three extraction routes: Claude Vision for complex layouts, Ollama + Qwen2.5-VL for private/local deployments, and Tesseract as fallback. Includes a /classify-image endpoint that tells you what's in a photo — useful for invoice processing, inventory, and auction-listing aggregators.
550-document benchmark, SROIE 95%.
VoiceFlow — record audio in the browser and get back structured output. Meeting notes with action items (owner, deadline, priority) or sales call analysis with CRM-ready JSON. Routes to Claude Sonnet 4.6 for sales calls where nuance matters, and Groq for speed on meeting notes.
RAGeval — self-hosted LLMOps observability for RAG systems. One decorator, SQLite by default, dashboard included. Scores retrieval relevance, groundedness via multi-judge consensus, faithfulness, cost, and latency. Persona-aware: it catches when your CFO response pulls data it shouldn't. Published to PyPI.
StreamPulse— real-time multi-source data pipeline with live WebSocket dashboard. Ingests from webhooks, Gmail, Sheets, CSV, APIs. Classifies records by domain automatically. First-class n8n integration — custom node template plus three importable workflows. Composes with DocIntel for vision-bearing payloads like auction listings.
Stack: Python · FastAPI · LangChain · LangGraph · PostgreSQL · ChromaDB · React · Docker · Claude · OpenAI · Groq · LiteLLM
LLMs: Claude Sonnet 4.6 · Claude Haiku 4.5 · GPT-4o / GPT-5 · Gemini 2.5 Pro · Groq LLaMA 3.3 · DeepSeek V3 · AWS Bedrock · Ollama (local)
Also: PyTorch · Scikit-learn · Pandas · faster-whisper · Tesseract · Prometheus · n8n · LiteLLM · OpenAI SDK
Open source: rageval (PyPI) · omnismart-personas (PyPI) · 6 GitHub repos
Languages: English (fluent) · French (native)
If you have a specific problem and a target outcome, send me a short brief. I'll respond with a structured plan, a realistic timeline, and a fixed-price estimate.
Steps for completing your project
After purchasing the project, send requirements so Yacine can start the project.
Delivery time starts when Yacine receives requirements from you.
Yacine works on your project following the steps below.
Revisions may occur after the delivery date.
Scoping call
We align on your data sources, executive roles, and deployment environment (cloud vs on-prem)
Data & auth setup
Connect your databases, configure JWT + RBAC roles, seed the ChromaDB knowledge base with your documents.