Hi, welcome to my site

Ivan Bosso

I'm a

I make your data speak, I streamline your processes with AI, take repetitive tasks off your plate, and give you time back.

As a Data / AI Engineer, I build machine learning, LLM and generative-AI solutions that actually reach production: RAG, fine-tuning, automation. From data engineering to full-stack development, I cover the whole chain to turn your data and processes into concrete results, at scale.

See my work Download my résumé

Problem

Data

Model

Production

Value

// What you can count on

What I actually bring you

On a client mission or inside your team: tools that genuinely get used, and an impact you can measure.

Real tools, not a prototype

I ship tools used in production, every day — not a demo that ends up forgotten in a corner.

An ROI you can measure

Time saved, fewer errors, lower costs: I quantify the impact. You know what it returns.

Autonomous, end to end

From data to interface, I cover the whole chain. Operational fast, with no need to spell everything out for me.

I fit into your domain

Industry, finance, payroll, real estate: I adapt to your field and your team.

Featured projects

Real cases: the problem, what I did, what changed.

View all →

CGR International

Applied AI

LLM extraction of technical data for industrial quoting

Read heterogeneous technical drawings with a vision LLM and extract structured data to automatically feed the quoting calculator.

In prod

AI document extraction

Case study →

ODERIS

Applied AI

AI classification of slides at scale for due diligence

Automatically file Vendor Due Diligence report slides into a business taxonomy, at the scale of 250,000+ slides, while keeping LLM call costs under control.

At scale

AI document classification

Case study →

IGAM

Data

Unsupervised NLP detection of recurring topics

An end-to-end unsupervised NLP pipeline to map the recurring topics in the inbound email flow of an accounting firm.

End to end

unsupervised NLP pipeline

Case study →

Valloire Habitat

Data

Multi-source territorial real-estate scoring

Built over 7 months a full data → decision chain assessing the real-estate appeal of a municipality.

7 months

data → decision chain

Case study →

// How I work

People before tools

Beyond the tech, my conviction: AI is a lever, not a goal. I use it to create value, where it genuinely makes sense.

People first

Beyond performance, I design for the people who will use it. A tool only has value if it's adopted.

Not everything should be automated

It's not a race to automate. Some decisions must stay human, and I know where AI genuinely helps.

AI is a means, not an end

I start from value, not hype. The right tech is the one that solves your problem, not the trendiest one.

Value-driven, not tool-obsessed

Tools change fast; locking myself into one would be a risk. I stay agnostic and focus on results that last.

// whoami

In plain words

I start from your need, design the solution, ship it, and measure what it saves you. Technology is only a means.

ivan_bosso.py

class IvanBosso:
    role       = "Data / AI Engineer"
    expertise  = ["ML", "LLM", "RAG", "fine-tuning"]
    i_do       = "embed AI into your processes"
    i_deliver  = [
        "time saved",
        "reliable decisions",
        "measurable results",
    ]
    obsession  = "your outcome, not the hype"

    def collaborate(self) -> str:
        return "let's talk about your needs"