01 · Overview
The opportunity
ParkWest is building an AI practice that sits between strategy, creative, and engineering — shipping assistants, content systems, and internal copilots for operators who need AI to actually move the business. We're hiring an AI Prompt Engineer to own prompt design, evaluation, and lifecycle across LLM-powered products. You'll work with strategists, designers, and engineers to turn fuzzy briefs into reliable, measurable AI systems — and set the bar for how we ship prompts, agents, and retrieval pipelines.
02 · A day in the life
What the week looks like
Monday is a discovery session with a client and our strategy lead — mapping a workflow we're about to automate. Tuesday you're writing and versioning prompts in our eval harness, running 200-case test suites against three model providers. Wednesday you pair with engineering on a retrieval pipeline and tune a router agent. Thursday is a craft review on a content assistant going live Friday, plus a regression run before deploy. End of week is a written postmortem on last week's launch and a calibration session with the team.
03 · What you'll own
The work itself
- 1Design, version, and document prompts and agent workflows across client and internal products.
- 2Build evaluation harnesses — golden sets, rubrics, automated graders — so quality is measurable, not vibes-based.
- 3Partner with engineering on retrieval pipelines, tool use, function calling, and guardrails.
- 4Benchmark model providers (OpenAI, Anthropic, Google, open-source) on cost, latency, and quality per use case.
- 5Translate client workflows into structured prompts, schemas, and fallbacks engineers can ship.
- 6Own prompt lifecycle: monitor production traces, debug regressions, and ship improvements weekly.
04 · What success looks like
How we'll measure it
01
AI features you ship have documented evals and a measurable quality bar, not anecdotal demos.
02
Production prompts improve in quality, cost, or latency quarter over quarter.
03
Engineering and strategy both trust your specs enough to build and sell from them.
05 · What you bring
The fit
Must-haves
- 2+ years working hands-on with LLMs in production — prompts, agents, RAG, or fine-tuning.
- Fluency with at least two major model providers and their tool-use / structured-output APIs.
- Strong written communication — your prompts and specs read like documentation, not notes.
- Comfort with Python or TypeScript for evals, scripts, and lightweight pipeline work.
- A portfolio or writeups of AI systems you've shipped, with honest discussion of failure modes.
Bonus points
- Experience with eval frameworks (Braintrust, Promptfoo, LangSmith, Ragas) and tracing tools.
- Background in linguistics, technical writing, or applied ML research.
- Experience shipping agentic systems with multi-step tool use in production.
06 · Compensation
What we pay, openly
Salary band
$110,000 – $140,000 / year
Final offer depends on experience, qualifications, and internal equity. We benchmark against the Canadian agency market each year and share banding openly. Reviews happen annually and at promotion.
07 · Benefits
Built to support the work
We invest in the team so the team can invest in the operators we work with. The full package, in plain language.
Competitive salary
Transparent banding, benchmarked against the U.S. agency market and reviewed annually.
Performance bonus
Annual bonus tied to company and personal goals — written down at the start of the year, not invented at the end.
Extended health & dental
Comprehensive medical, dental, and vision coverage for you and your dependents from day one.
Mental health support
Confidential EAP plus a dedicated mental-health benefit for therapy, coaching, and crisis support.
Flexible PTO + paid sick days
Generous paid time off plus separate paid sick days so you can rest without trade-offs.
13 paid holidays
All statutory holidays plus your birthday and a floating day to use whenever it matters.
Hybrid by design
Two anchor days a week in the studio, the rest of the week wherever you do your best work.
Home office stipend
A one-time setup stipend plus a monthly allowance for tools, internet, and workspace.
Professional development
Annual learning budget for conferences, courses, books, and certifications.
Paid volunteer time
Paid hours every month to contribute to causes and communities you care about.
RRSP matching
Employer matching on your retirement contributions from your first day.
Parental leave top-up
Top-up on Employment Insurance benefits for new parents — for any path to parenthood.
08 · How we hire
What to expect
- 01
Apply with your work
Send a note, a portfolio or relevant samples, and a short read on why this role fits. We read every message.
- 02
Intro conversation
A 30-minute call with the hiring lead to talk through your background, the role, and your questions.
- 03
Craft session
A working session — paid where it requires meaningful prep — designed to mirror real work, not a brain teaser.
- 04
Meet the team
Final round with the people you'd work with day to day, plus a leadership conversation on growth and trajectory.
Ready to apply?
Send us a note and your best work.
Tell us why this role fits, what you'd want to own first, and a couple of links to work you're proud of. We read every message.
Apply for AI Prompt Engineer