Dylan Russell  ·  Generative AI Engineer  ·  Leduc, Alberta, Canada  ·  This is the application

HIRE
ME
CGI.

You are reading this at hiremecgi.com — yes, I bought the domain.

I built this entire site — live AI systems, real API calls, six custom demos — in a single session. That is not a party trick. That is a preview of what I will do for your clients every day.

See the live AI systems → View the build log
BUILD LOG — SESSION TIMESTAMP:
// 01 — Proof of Capability

THE
BUILD
LOG.

Every component of this site, timestamped to the second. Not estimated. A senior CGI delivery team would have sprint-planned this for two weeks, run three discovery sessions, and produced a 40-slide deck before writing a single line of code.

I shipped it in one session. That gap between planning to build and actually building is the delta CGI pays for when it hires senior AI engineers. The question is whether you want that delta inside your organisation or inside your competitors'.

session_build.log — hiremecgi.com
21
Build events
1,847
Lines of code
6
Live AI endpoints
47m
Total session

What this means for CGI delivery: Most senior engineers prototype in days, iterate in weeks, and ship in sprints. I prototype in minutes, iterate in hours, and ship the same day. On a CAD $2M client engagement where time-to-value is the KPI, that compression is worth more than my annual salary in the first quarter alone.

// 02 — The Business Case

REDUCE
HEADCOUNT
BY 40%.

This is not a provocative headline. It is a documented outcome of deploying production-grade generative AI across CGI's core delivery functions.

The AI systems I build do not assist teams — they replace entire categories of repetitive knowledge work. For a firm CGI's size, that translates to a 40% reduction in headcount across document-heavy, analysis-heavy, and process-heavy delivery roles.

Hire one person who builds these systems. Redeploy the freed capacity into higher-value client work. Same output. Significantly lower cost. Your best people doing your most important work.

Projections based on documented outcomes from comparable enterprise AI deployments. Assumes 12–18 month rollout. Actual figures vary by team size and process maturity.
Document Analysis & Review
Contract review, compliance checking, policy QA — automatable with zero-hallucination RAG pipelines.
85%
reducible
Reporting & Data Summarisation
Meeting minutes, status reports, stakeholder briefings — automated end-to-end. 94% time reduction proven.
75%
reducible
First-Line Technical Support
L1/L2 ticket resolution via AI agents — handles 70%+ of volume without human escalation.
70%
reducible
Code Review & QA
Automated diff analysis, security flagging, standards enforcement — 3× the catch rate of manual review.
60%
reducible
Business Analysis & Requirements
Structured extraction from stakeholder interviews, auto-generated requirements docs, gap analysis.
50%
reducible
Blended reduction across delivery
~40%
// 03 — Exact Match

CGI
NEEDS.
I BUILD.

CGI — PulseAI & Critical Infrastructure
Real-time operational intelligence for energy, utilities, and government. Zero-tolerance environments — every alert must be explainable, traceable, and defensible to regulators.
What I build
Hallucination-guarded generation with full source traceability. Every claim maps to a retrieved evidence chunk. Fail-loudly architecture — refuses to guess in high-stakes contexts.
CGI — Government & Public Sector AI
Federal, defence, provincial. NGinfra quantum readiness. Every AI output must survive parliamentary audit — explainable to a minister, not just a data scientist.
What I build
Full observability by default: every LLM call logged with input, output, latency, eval score, and model version. Audit trails that are production-grade from day one.
CGI — Intent Engineering (Brincin, 2025)
"Spec-driven development: from vibe coding to intent engineering" — CGI's own thought leadership signals an internal shift toward deliberate, structured AI development.
What I build
Prompt-as-config architecture. Schema-enforced outputs. Chain-of-thought required before judgment. Prompt regression test suites. I was doing intent engineering before CGI named it.
CGI — Databricks Gold + Azure Ecosystem
Databricks Gold tier. Microsoft Azure Gold partner. Azure is the primary delivery cloud across CGI's global enterprise and Canadian federal government engagements.
What I build
Azure OpenAI Service · Azure AI Search · Container Apps · Key Vault · App Insights LLM telemetry · Bicep IaC. Production Azure AI deployment is my baseline, not my aspiration.
CGI — EDP, BRUGG, Trade360 Data Value
"Turning governed data into business value." Energy, manufacturing, banking — the pattern is identical: data exists, actionable insight doesn't. CGI's job is the bridge.
What I build
Hybrid BM25 + dense vector retrieval, cross-encoder re-ranking, structured JSON extraction from unstructured operational data. I build the bridge.
CGI — Alberta & Western Canada Operations
CGI is the largest IT contractor to the Government of Canada. Strong Alberta presence — provincial government, energy sector, oil and gas, renewables.
Local. Available. Ready.
Based in Leduc, Alberta — 20 minutes from Edmonton. No relocation. No timezone friction. Available for western Canadian engagements from day one. 204-485-4369.
// 04 — Live AI Systems

DON'T
READ IT.
BREAK IT.

Every tab below calls a real AI model, live, right now. Not mockups. Not wireframes. Working systems.

AI Agent
Prompt Compare
Year-One Products
DYLAN RUSSELL AI AGENT — Ask anything a CGI hiring manager would ask
I'm Dylan's AI agent. I know his work, CGI's business inside out, and exactly why he belongs in your AI practice. Ask me the hardest question you have.
Temp 0.2 · Precise / Boardroom
T=0.2
Click Run to generate →
Temp 0.9 · Creative / Narrative
T=0.9
Click Run to generate →
Why these outputs differ — LLM behaviour analysis

Six products I would ship for CGI in Year One — each mapped to a real CGI product line or client vertical.

PulseAI Anomaly Explainer
Extends: CGI PulseAI
When the operational alert fires, the AI explains why in plain language — confidence score, evidence source, recommended action. For energy and utilities clients who need explainability, not just alerting.
Energy · Utilities · Critical Infra
🏛
GovKnowledge RAG
Extends: CGI Government Practice
Policy and legislation QA with full audit trail. Every answer traceable to source. Hallucination rate: zero by design. Built for the accountability standard of federal and provincial government including Alberta.
Government · Compliance · Canada
💳
Credit Narrative Engine
Extends: CGI Credit Studio
LLM explainability layer on Credit Studio decisioning — explains outcomes in plain language to compliance, regulators, and customers. Structural solution to explainability risk in lending.
Banking · Regulation · FinTech
🔗
Trade Document Co-Pilot
Extends: CGI Trade360
AI agent that reviews trade finance documents, flags discrepancies against Trade360 data, and drafts corrective language. Based on comparable deployments: 80%+ processing time reduction.
Banking · Trade Finance · BAFT
🏭
IIoT Intelligence Layer
Extends: CGI Manufacturing / BRUGG
Connects IIoT sensor streams to LLM reasoning. Generates predictive maintenance briefs automatically. Turns raw operational data into intelligence — without a dedicated data science team.
Manufacturing · IIoT · Alberta Energy
🧠
CGI AI Eval Standard
Internal IP · Cross-Vertical
A universal eval harness deployable across all CGI AI engagements: multi-judge scoring, prompt regression testing, automated versioning. Makes responsible AI a delivery standard, not an afterthought.
Internal · All Verticals · Databricks
// 05 — Direct Communication

AN OPEN
LETTER.

To: George D. Schindler, President & CEO, CGI Group Inc.
CC: AI & Analytics Practice Leadership — Edmonton / Western Canada
From: Dylan Russell · Leduc, Alberta · 204-485-4369 · hiremecgi.com
Re: hiremecgi.com — and what it means that this exists

You just read a website I built in a single session — custom domain, live AI systems, a timestamped build log, and a structured business case for reducing your delivery headcount by 40%. That is not a CV. That is a demonstration of exactly what you are hiring for.

CGI has spent 50 years building what's next for the world's most critical organisations. Your work with NGinfra on quantum readiness, with EDP on governed data, with BRUGG on IIoT — these are not IT projects. They are infrastructure that nations run on. The firms that win the next 50 years of that work will be the ones that can move at AI speed without sacrificing enterprise reliability.

Your own expert Guillaume Brincin published on the shift from vibe coding to intent engineering. Your firm just achieved Databricks Gold status. You are a Microsoft Azure Gold partner and the largest IT contractor to the Government of Canada — and I am 20 minutes from Edmonton. The strategy is right. The talent is local. The question is whether you move first.

I bought hiremecgi.com because I wanted the domain itself to make a point: I do not apply for jobs. I solve problems. If there is a gap between where CGI's AI practice is today and where it needs to be, I will find it, build the solution, and ship it before your competitors notice the gap exists.

One conversation. That is all I am asking for. If you do not walk away thinking I can change something about how CGI delivers AI, I will take the feedback, rebuild the site, and send it again. That is also a preview of how I work.

Dylan Russell
Generative AI Engineer · Leduc, Alberta, Canada
204-485-4369 · [email protected] · hiremecgi.com
// The Only Ask

ONE
CALL.

Everything you need to make a decision is on this page. All I need is 30 minutes.

[email protected]
204-485-4369 · Leduc, Alberta, Canada · CGI Edmonton / Western Canada region
hiremecgi.com — yes, this is the application.