Skip to content
View salvo8519's full-sized avatar

Block or report salvo8519

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
salvo8519/README.md

Salvatore Miseo

AI developer. I bring AI into how companies actually work, and take it to production: agentic pipelines, RAG systems, AI gateways and integrations with ERP, WMS and TMS.

I start from the process. As a functional analyst I redesigned business processes for digital in more than 20 client projects. Before that, I spent a decade in finance and real estate: credit analysis for SMEs, risk and valuation of a property portfolio worth around 5 billion euro, investment analysis, business plans for complex developments. That background shapes how I build software: results must be exact, traceable and worth what they cost.

What I work on

  • Agentic pipelines in production - a multi-tenant platform that reconciles supplier invoices, orders and proofs of delivery against ERP costs, with LLMs as the decision engine and a human in the loop.
  • Enterprise integrations - APIs and connectors towards ERP, WMS and TMS, built on versioned data contracts, layers and adapters.
  • Document AI - agentic extraction from complex documents; research for a project funded by the iNEST programme (PNRR) on AI for document management.

I work spec-driven and contract-first, in sprints.

Private projects

AI workflow orchestration platform - in progress, moving towards maturity. Automation workflows served as enterprise-grade APIs: a workflow enters a registry and becomes an authenticated endpoint, with grants, multi-level quotas, durable executions and retries. Contracts on input, output and lifecycle; usage metered in units of work for billing; audit and GDPR pseudonymisation; white-label per company. The automation engine is a configuration choice, plugged in through adapters. Next: structured document parsing, and personal-data anonymisation, which I plan to release as a separate open-source service written in Rust.

Enterprise help desk with RAG - multi-tenant knowledge platform for companies and departments, with role-based access, a knowledge base hub and an AI help desk agent. FastAPI, React. Working MVP; next step is moving the vector store to Qdrant.

Enterprise authentication service - designed together with its security, architecture and GDPR analyses, including a record of processing activities. FastAPI.

Open source

Fairmark - Deterministic valuation engine for analysts and AI agents. The model reads and writes, Fairmark does the math: exact, reproducible, traceable. Rust core, Python SDK, MCP. Docs · PyPI · crates.io

Wayloop - Spec-driven agentic loop for whole software projects: from idea to merged code, in your issues, with humans at the gates. Rust core. Pre-alpha. Docs

Stack

Python (FastAPI, asyncio) · Rust · PostgreSQL · Redis · Celery / RabbitMQ · Docker · LangGraph · LangChain · LiteLLM · Langfuse · Qdrant / pgvector · MCP

Background

Master in Real Estate Finance, SDA Bocconi – MIP Politecnico di Milano · Degree in Economics and Business Administration, University of Salerno

Contact

LinkedIn · Salerno, Italy · Italian, English

Pinned Loading

  1. fairmark fairmark Public

    Deterministic valuation engine for analysts and AI agents: exact decimal arithmetic, full audit trail. Rust core with Python SDK, CLI, MCP server and function-calling tools.

    Rust 1

  2. wayloop wayloop Public

    Spec-driven agentic loop. From idea to merged code, in your issues, with humans at the gates. Rust core, open source.

    Rust