
Senior ML Engineer (Client Solutions)
Posted Aug 31, 2026
The role in short
Build and deploy production ML forecasting and detection models directly inside client environments, handling the full lifecycle from data discovery to monitoring.
About AZX
Our mission is to accelerate positive impact in critical industries through AI transformation. We specialize in physics-informed ML and enterprise AI solutions that directly address climate and sustainability challenges.
We’re growing quickly and already work with category-leaders in real estate (CBRE), energy (LevelTen Energy), logistics (Flexe) and utilities.
We bootstrapped profitably for our first year and are now backed by leading investors focused on AI, climate and energy.
We work on challenges in clean energy, decarbonization, climate risk, energy systems, and global economics. We’re building our company for long-term success and aim to create the ultimate place to work for those passionate about AI and making a positive impact.
About This Role:
We are seeking an ML Engineer who builds ML systems directly inside client environments. Your job starts with the client's actual data spread across multiple systems — and ends with a model running on a schedule inside their environment. You will bring a strong area of expertise, but expect to wear many hats as part of a small team — some DevOps, some infrastructure, some front end and back end — because you are the engineering face of AZX to your client.
Responsibilities:
- Own the full ML delivery lifecycle: data discovery and cleaning, modeling, evaluation, deployment into the client environment, scheduling, monitoring, and retraining policy.
- Build forecasting and detection models that hold up against real-world data quality issues (late feeds, revised rows, missing labels).
- Backtest and evaluate models honestly enough to stake real operational decisions on them, and defend your precision/recall tradeoffs to the people who bear the cost of false alarms.
- Design systems that distinguish "no prediction" from "wrong prediction," so a missing answer reads differently to the end user than an incorrect one.
- Ship enough product to make the model usable — a FastAPI service, a small React surface, a scheduled job — whatever "usable capability" means for that client.
- Own the measurement story: agree on baselines and KPIs before deployment, instrument for monitoring, and deliver a post-deployment readout with attribution limits clearly stated.
- Maintain client-facing engineering presence and a feedback loop into the platform team — running discovery, working sessions with client IT/data teams, demos, and surfacing the data shapes and failure modes only visible from inside client data.
Core Qualifications:
- 5+ years of shipping applied machine learning to production — forecasting, detection/classification on time series, survival/reliability modeling, or optimization — with an evaluation you defended to someone whose job depended on it.
- Strong data engineering skills and willingness to use them: you find, clean, join, and profile data yourself at awkward scale, without a dedicated data team.
- Rigorous validation discipline — chronological splits, walk-forward validation, as-of correctness, and an instinct to be suspicious of a suspiciously good metric.
- Enough software engineering to ship real systems: Python, SQL, tests, Docker, a scheduler, an API or app surface, and monitoring — type-strict, tested, reviewable code, even in a pod of two.
- Client-facing capability and the assertion to use it — running discovery, leading demos, and pushing back early and plainly when an ask is wrong, with an alternative already in hand.
- Judgment about when ML is the wrong tool, and the willingness to say so to a client who wants AI regardless.
- Practical fluency with our core stack — Python 3.12+ (pandas/polars/DuckDB, scikit-learn, statsmodels, gradient boosting), SQL/Postgres (with TimescaleDB/PostGIS for grid work), and time-series feature engineering and validation.
- Comfort building the surfaces that make a model usable — FastAPI plus enough React/TypeScript to expose results — and deploying it with Docker and basic cloud tooling (Azure/AWS).
- Working fluency with LLMs for the agentic edges of client work (extraction, retrieval) — depth isn't required, but honesty about your actual experience is.
- Bachelor's Degree; Master's is a Plus
- Domain experience in Energy, Utilities, Infrastructure, and Commercial Real Estate is a plus
Why AZX!
- Be part of a fast-growing, profitable, mission-driven company with industry-leading clients tackling the massive opportunity of AI transformation in critical industries.
- Competitive early-stage startup compensation (based on capabilities, experience, and location)
- Bonus eligibility
- Health insurance with meaningful coverage for dependents
- Flexible paid time off
- Equity
- Fully remote culture with a cluster of teammates in Seattle
Additional Information:
- Must be able to travel 2x/year for company summits
- Applicants must be currently authorized to work in the United States on a full-time basis.
- We are unable to sponsor or take over sponsorship of employment visas at this time.
- Please note that our interview process includes a written take-home assignment followed by a live two-hour technical session with our engineering team, so if that format isn't a good fit, we'd ask that you not apply
- Please only apply to a maximum of 2 roles at a time, any applicants who apply to more then 2 roles within a 6 month period will automatically be disqualified
Next Steps:
If this job sounds like a great fit but you don’t check ALL of these qualification boxes, we’d still love to hear from you!
Prep for this interview
5 questions likely to come up at AZX, each matched to a line in this posting.
- 1Cloud, deployment and security
They wrote: “Comfort building the surfaces that make a model usable — FastAPI plus enough React/TypeScript to expose results — and deploying it with Docker and basic cloud tooling (Azure/AWS).”
How would you debug a deployment that works in staging but fails at the customer's site?
How to answer
Why they ask: Customer environments differ in ways you can't see. This tests method under pressure.
- Compare versions, config and environment variables
- Check network rules, proxies and certificates (common in enterprise)
- Look at resource limits and permissions
- Add the missing check to your install steps afterwards
Tip: List the usual suspects and how you'd check each without full access.
- 2LLMs, RAG and agents
They wrote: “Working fluency with LLMs for the agentic edges of client work (extraction, retrieval) — depth isn't required, but honesty about your actual experience is.”
A customer wants a chatbot over their internal documents. How do you build it?
How to answer
Why they ask: RAG is the most common first project. They want the whole path, from messy files to answers people trust.
- Collect and clean the documents; handle permissions per user
- Split them into chunks, embed, and store with source links
- Retrieve, then answer with citations; say “I don't know” when retrieval is weak
- Build a test set of real questions with expected answers before launch
Tip: Spend more time on the data and on testing than on the model.
- 3Discovery and scoping
They wrote: “Own the full ML delivery lifecycle: data discovery and cleaning, modeling, evaluation, deployment into the client environment, scheduling, monitoring, and retraining policy.”
A customer says “we want AI in our claims process.” What do you ask in the first meeting?
How to answer
Why they ask: Most FDE work starts this vague. They want to hear you find the real problem before you pick a tool.
- Map the current process: who touches a claim, in what order, with which systems
- Find the slowest or most error-prone step and ask what it costs them
- Agree on one measurable goal (for example time per claim, or error rate)
- Ask what data you can access, and who signs off on security
Tip: Ask about the work, not the technology. Who does it today, how long it takes, what a mistake costs, and how they'd know it got better.
- 4Integrations and APIs
They wrote: “Enough software engineering to ship real systems: Python, SQL, tests, Docker, a scheduler, an API or app surface, and monitoring — type-strict, tested, reviewable code, even in a pod of two.”
A customer's API starts returning errors in production at 2am. What do you do?
How to answer
Why they ask: FDEs often own production for their customers. This tests incident handling and communication.
- Check your side first: logs, recent deploys, retries piling up
- Stop the damage (pause jobs, queue requests, turn off the feature)
- Tell the customer what you see and what you've done, in plain words
- Afterwards: add a check that would have caught it, and write a short report
Tip: Split the answer into the first hour and the week after.
- 5Travel
They wrote: “Must be able to travel 2x/year for company summits”
How do you handle being on the road with customers for weeks at a time?
How to answer
Why they ask: Many FDE roles involve regular travel. They check you know what you're signing up for.
- Your past travel or on-site experience
- How you stay productive away from the team
- What travel level works for you, said plainly
Tip: Be honest about what works for you. A mismatch here hurts both sides later.
Not the one? Get the next one by email
New roles land here every day. Pick what you want to hear about.
Want the job, not just the posting? Learn how FDEs work.
FDE Info is our free weekly newsletter for forward deployed engineers: tools, field notes from real deployments, and how FDE teams hire and pay.
Free. One email a week. Unsubscribe in one click.