Philippe Tremblay

Montréal, Québec

Philippe Tremblay

Distributed systems and cloud infrastructure — a decade of it, at Nasdaq, Metrio, Element AI and Desjardins. I take a small number of consulting mandates a year, mostly putting coding agents to work under real engineering discipline.

Practice

01
01

Agent harness enablement

Coding agents in a real repository, without losing control of it: instructions the agent actually follows, hooks that enforce process rather than suggest it, gates that fail closed, and a register that makes unfinished work visible instead of invisible. Delivered as a two-day team workshop, or as implementation alongside your engineers.

02

Private and sovereign AI deployment

Model routing that keeps sensitive turns in-region, private inference on Kubernetes or OpenShift, and the cross-border transfer assessments Québec's Law 25 requires — which almost nobody has actually written.

03

Platform and reliability engineering

Golden paths, infrastructure as code, deployment pipelines, and enough observability to know whether any of it is working.

04

Technical due diligence

Scalability, security posture and IP hygiene, read the way a site reliability engineer reads them. For acquirers, and for founders about to become one.

05

Observability migration

OpenTelemetry, Prometheus and Grafana — usually replacing a per-host bill that stopped making sense two renewals ago.

Currently

02

Co-founder of Dravr, a multi-tenant platform written in Rust with a production MCP server across three transports. Two people build it, running the harness described above every day. Most of what is on this page comes from there.

Credentials

03
01

Certified Kubernetes Administrator

CKA
02

Google Cloud — Professional Cloud Architect

GCP
03

Google Cloud — Professional Data Engineer

GCP
04

McGill University — Software Engineering

B.Sc.

Off the clock

04

Racing with Messorem Club Cycliste. Before that, seven years on the ITU triathlon circuit.