the agent

Ask Othmane's agent

It answers from the résumé and the stories behind it, in your language, and can take a note for him if you want him to get back to you. Type, or press the microphone and say it.

conversations are recorded · sessions expire after a day

full-stack · ai · forward-deployed · software engineer

OTHMANE
ZIZI_

I build AI products end-to-end — React/TypeScript interfaces, FastAPI & Flask backends, data pipelines, LLM agents — and I ship them from inside the field: embedded with aerospace buyers, web3 CFOs, and market researchers on two continents. The stack doesn't end at infra. It ends at the customer.

montréal, qc · utc−5 · full-stack ai forward-deployed engineer · deployable anywhere · portfolio →

the console

fdx — forward deployment console

the stack

Full-stack, with receipts

Six layers, top to bottom. Tap a layer to filter the deployment log below to the roles where I actually worked in it.

deployment log

Seven deployments, reverse-chronological

Expand any record for the mission log.

Senior AI Engineer @ Matador AI

2026-09 → now

Conversational AI voice agents for car dealerships across North America. Building Conductor, the harness for all internal and customer workflows, and the agents that run on it.

▸ expand log
  • Building Conductor, the harness for all internal and customer workflows, and on it proactive, long-running, personal AI agents driving sales and service for car dealerships across North America.
  • Building Call AI, Matador's voice AI agents product.
servicesmodelsinfra 45.50°N 73.57°W · montréal

AI Infrastructure Lead – Master's Internship @ Meroka

2026-01 → 2026-06

8VC-backed, AI-powered private equity. Owned end-to-end engineering of an AI-native platform — discovery & roadmap through implementation, model evaluation, infrastructure, and the UX for LLM-powered workflows.

▸ expand log
  • Owned the platform end-to-end — product discovery & roadmap through implementation and infrastructure.
  • Shaped model evaluation — deciding what "good" means before shipping it.
  • Built in direct collaboration with early users — the forward-deployed loop, compressed to days.
  • Built a multi-agent enrichment system that qualifies every lead and person the moment it enters the CRM — specialist agents fired in parallel by a Temporal orchestrator on Bedrock, failures re-run with the successes as context.
  • Shipped a medallion data pipeline (Bronze → Iceberg Silver → Spark) with EventBridge orchestration and full lineage — hundreds of terabytes of insurer transparency data, rebuildable from scratch with one command.
interfaceservicesmodelsinfrafield 45.50°N 73.57°W · montréal

Data Architect (Data Engineering Scope) – Master's Co-op @ CAE

2025-10 → 2026-06

Master's co-op inside one of the world's largest aerospace & flight-simulation manufacturers — ML for procurement.

▸ expand log
  • Led data engineering and ML initiatives to optimize aerospace procurement operations.
  • Designed pipelines and models that surface actionable insight across a multi-source supply chain.
datamodels 45.50°N 73.57°W · montréal

Business Analyst (Software Engineering Scope) @ Integral

2024-10 → 2025-02

Electric Capital-backed web3 financial-operations platform. Backend and data engineering for enterprise CFOs.

▸ expand log
  • Engineered the core product's backend notification system, serving web3 CFOs across enterprise customers.
  • Shipped an automated billing & reconciliation system; built internal tools that automated large parts of support.
  • Delivered data engineering for new blockchain and accounting integrations.
servicesdata 45.50°N 73.57°W · montréal

Junior Consultant (w/ Software Engineering Scope) @ AMS Africa

2022-10 → 2024-06

Real-estate strategy & technology. The full-stack proof: one person, all six layers, deployed at the client.

▸ expand log
  • Built and deployed an AI knowledge-management platform into production that automated core market-research workflows.
  • Designed a custom retrieval-augmented agent trained on proprietary company data — before it was a product category.
  • Owned consulting engagements end-to-end. Stack: React, Flask, PostgreSQL, Google Cloud.
interfaceservicesdatamodelsinfrafield 33.57°N 7.59°W · casablanca

Director of Product Management @ Finny (acq. by Origin)

2022-02 → 2022-05

Crypto & fintech software, San Francisco Bay Area. Led the product team through shipped roadmap.

▸ expand log
  • Shipped onboarding automation, an A/B-test-validated payment option, and an international version of the product.
  • Published and edited crypto education and The Gist, a finance & investment newsletter.
interfacefield 37.77°N 122.42°W · sf bay area

Content Creator (Business Research & Audience Development) @ Thinknum

2021-10 → 2022-02

Alternative data, New York. Web scraping and dataviz on the inside, enterprise pre-sales on the outside.

▸ expand log
  • Applied advanced web scraping and data-visualization to author crypto research and generate qualified leads.
  • Led pre-sales calls and SaaS demos — translating complex data products for enterprise clients.
datafield 40.71°N 74.01°W · new york

pretraining

Four schools, three countries, three scholarships

Before the deployments came the training runs. Each one crossed a border.

  1. master

    McGill University · Desautels Faculty of Management

    Master of Management in Analytics

    Montréal · the analytics master's, taken while shipping in production at CAE and Meroka.

  2. bachelor

    Minerva University

    Bachelor of Arts & Sciences · Business, Arts and Literature

    San Francisco · the world's most selective college program, ranked #1 worldwide for innovation; a degree taught in residence, city by city, on four continents.

  3. pre-university

    UWC-USA

    International Baccalaureate · Higher Level: Mathematics, History, Music

    Montezuma, New Mexico · scholarship from Morocco to a United World College, the movement presided over by Nelson Mandela, King Charles III and Queen Noor.

  4. origin

    Lycée Lyautey

    French Baccalauréat · scientific track, further mathematics

    Casablanca · the French lycée where the trilingual default was set.

model card

othmane-zizi · human-native, agent-fluent

Specifications of the engineer being deployed.

modalities
french · arabic · english (native ×3) · spanish (professional)
training distribution
lived and thrived on five continents — distribution shift is the home domain
post-training
McGill MMA · Minerva University · UWC-USA · Lycée Lyautey — the full runs are under pretraining
evals
Databricks AI Agents Hackathon — 2nd place · Asset Management Hackathon — finalist
certifications
IBM Full Stack Software Developer · Python for Data Science, AI & Development · Statistics & R
known behaviors
ships end-to-end · embeds with users · runs own pipelines as funnels with measured conversion
context format
/resume.md — plain markdown, built to be read by your agent
cold-start latency
none — currently deployed in production at Matador AI

endpoints

Open a connection

All endpoints are live. The humans answer in the language of your choice; the markdown answers your agent.

GET/portfolio eight projects told one screen at a time, then every public repository POST/email othmane.zizi.pro@gmail.com GET/linkedin linkedin.com/in/othmanezizi GET/github github.com/othmane-zizi-pro GET/resume.md plain markdown — download it, drop it in your agent, ask it anything GET/resume.pdf the classic one-pager, for your ATS