Job Description

Powell Industries is in the middle of a meaningful shift in how we use data to run the business. Over the years, our teams have developed genuine ingenuity to make data work for them: curating Excel exports, building secondary processes to transform and distribute information, and creating workarounds that keep operations running even when the underlying systems were not designed with that in mind.

What we are doing now is taking that same ingenuity and re-baselining it on a modern platform — one where those workarounds can be de-siloed, prototyped freshly, and properly collaborated upon, enabled by a central team focused on maintaining a meaningful balance between delivery and enabling democratization of delivery. The institutional knowledge embedded in those Excel files and secondary processes does not disappear; it becomes the starting point for something governed, shared, and built to last. The time to ramp up on this new approach has come, and with it, the steps to move on from the legacy tools that have carried us this far.

The Data Analyst sits at the center of that transition — helping translate business questions into trusted data products, accelerating the shift from report-pulling to genuine self-serve analytics, and compressing the time between a question and a confident answer. Critically, that means making insights actionable within the data experience itself, not as a downstream step. And where possible, closing the loop upstream too: bringing data entry and collection into the same coherent surface.

Making data accessible and actionable

• Own and evolve the semantic data models and governed views that serve as the trusted foundation for workbooks, apps, and workflows built across the business — ensuring that what gets built on top is built on something solid, consistent, and shared.

• Enable builders, collectors, and reviewers across the business to work independently in Sigma — not by building for them, but by ensuring the underlying data model is well-defined, well-documented, and expressive enough that they can construct what they need without hitting a wall. Where AI tools can shorten that path — through prompt-driven query generation, notebook-based analysis, or faster pattern recognition — use them; the goal is a faster answer, not a showcase of the tooling.

• Curate and maintain common definitions, reusable metrics, and shared business logic so that teams across the organization are working from the same numbers — regardless of which surface, workbook, or workflow they are using to access them.

• Where the business has questions that structured data alone cannot answer — documents, emails, field notes, unstructured inputs of any kind — our underlying data platform unlocks the ability to bring that context into the analytical surface, alongside direct access to AI models. This is not a standing deliverable; it moves at the pace of genuine demand.

Distinguishing analytics from operational reporting

• Help the organization develop sharper instincts about the difference between operational metrics (what happened, is it in spec, does it need action today) and strategic analytics (why is it happening, what should we do about it). This distinction shapes what gets built, how fast it needs to update, and who needs to see it.

• Work iteratively with business teams to understand what they are actually asking, not just what they think they can request from a data system — and translate that understanding back into better models, cleaner definitions, and sharper questions.

Automating the repeatable

• Identify manual, recurring data workflows — the Monday-morning Excel refresh, the end-of-month report, the ad hoc pivot rebuilt from scratch every quarter — and replace them with automated, auditable pipelines that run without anyone asking.

• Identify patterns in the most frequently asked business questions and build reusable, parameterized notebooks and templates that make expert-quality analysis available without requiring an expert every time.

• Write clean SQL and Python to support automation, data validation, and lightweight transformation work on top of our cloud data platform.

You will need

• Solid SQL — comfortable querying, filtering, joining, and aggregating in a cloud data warehouse environment. Familiarity with how modern data platforms are structured matters more than experience with any specific one.

• Working Python — enough to write scripts, automate data tasks, and work confidently in notebook environments (Jupyter, Marimo, or similar).

• An instinct for clarity — you ask what a number means before you put it in a chart, and you notice when a metric definition creates more confusion than it resolves.

• Comfort with ambiguity — business questions rarely arrive fully formed, and you are good at turning a vague ask into a concrete, answerable question.

• Curiosity about AI tooling — you have used LLM-based tools to accelerate your own work and are interested in how they can make data accessible to people who do not think of themselves as data people.

Nice to have

• Experience with Sigma Computing or a comparable modern BI platform (Tableau, Looker, Power BI).

• Exposure to dbt or semantic modelling concepts.

• Any experience in a manufacturing, industrial, or operations-heavy environment — you will understand the data faster.

• Familiarity with pipeline orchestration or workflow automation tools.