Data Scientist (Python, AWS)

Job Description

Job Description

Peregrine Advisors is a firm founded on a simple conviction: the best solutions come from the people closest to the problem, given real ownership and the tools to deliver. We are a data and technology innovation hub and a Benefit Corporation working at the center of the federal government's mission to deliver for client stakeholders and the US public, looking for highly motivated contributors who thrive when trusted to own a hard problem and equipped to deliver the solution.

Your first assignment

Your first assignment will likely include transforming client data – structured, semi-structured, and unstructured – into analytic processes and outcomes that help drive mission-critical initiatives:

• framing analytical questions with the people who own the mission, and turning them into designs that can actually be tested;
• building analyses and models in Python and SQL using cloud-native data platforms hosted in AWS or Azure;
• designing the evaluation that provides the evidence a model is fit for a client environment: baselines, error analysis, and documented limits;
• hardening one-off notebooks into repeatable, reviewed analytical pipelines; or
• putting findings in front of decision-makers in plain language, with visuals that carry the argument instead of decorating it.

Adaptability and a demonstrated capacity to learn new concepts, systems, and tools are what keep you ahead of program needs as projects complete and priorities shift.

The work is real, hard, and it matters. It is also where you start, not the shape of your career here: we hire people, not seats, and we move our best to where the hardest problems are.

What you'll build

• Models that earn their place: forecasting, classification, risk scoring, and anomaly detection matched to the question, the data, and the scrutiny a federal decision has to survive.
• The evaluation evidence behind them: validation designs, error analyses, and documentation that lets someone else verify your analyses and outcomes.
• Analytical features and datasets drawn from large, messy, document-heavy corpora, including text extracted with Amazon Textract and analyzed with foundation models through Amazon Bedrock.
• AI-assisted analysis as standard practice: language models and coding assistants across the lifecycle, with your judgment on top.
• In time, the firm itself: new capabilities, tools, and lines of business you help spin up.

Who you are

You are rigorous about method and honest about uncertainty: when the evidence supports a narrower claim, you make the narrower claim. You care what the data is for, not just what it contains, and you would rather spend a day understanding the process that produced a number than a week modeling it wrong. You communicate clearly, adapt across roles and clients, and care about creating public value. You experiment, fail, learn, and repeat quickly. You would rather own an outcome than be handed a task.

What you bring

• A working command of statistics and machine learning applied to real operational data, not just curated coursework sets.
• The engineering habits that make analysis trustworthy: version control, reproducible pipelines, tested transformations, and documentation someone else can follow.
• The communication craft to make a technical result understandable to the people who depend on it.

The precision to work where accuracy and auditability are not optional.

What you'll need

Sole United States citizenship and the ability to obtain a Public Trust determination are required for this initial engagement. This is a hybrid role based in the Washington, DC metropolitan area, and it requires commuting into DC regularly. Everything else, the years, the certifications, the specific tools, we ask in the application and get into during the interview.

What we offer

A high-performing team of developers, engineers, data scientists, architects, and strategists solving complex, real-world problems, with work that runs from strategy formulation to hands-on implementation. We develop people across roles and clients, with extensive onboarding and sponsored training and professional development. And Peregrine has been a Benefit Corporation from day one: public value is built into the work itself, not bolted on afterward. Work worth your best years.

What we commit to

As a Benefit Corporation, our commitment runs three ways: real, measurable value for our clients; government that works better for the public; and a team that makes everyone in it better.

We hire people who want to help build the firm, not just work at it. If that is you, apply.

Peregrine Advisors is an equal opportunity employer.

Peregrine exclusively works with OPEN Data Jobs to recruit our team. Register with OPEN Data Jobs to be considered for this opening and future roles.

Requirements

• 5+ years of applied data science or statistical modeling experience
• Bachelor's degree
• Strong working command of Python and SQL
• Working knowledge of statistics, machine learning, and model validation
• Familiarity with cloud-based data and analytics services on Amazon Web Services (AWS)
• Ability to carry an analysis from question to defensible, documented answer
• Basic proficiency in writing, PowerPoint, and Excel

Preferred

• Master's degree in a relevant field
• Experience with modern data tools and platform (Amazon Aurora, S3 with Apache Iceberg, AWS Glue, or Trino)
• Experience applying large language models to analytical work (Amazon Bedrock or comparable services)
• Data-visualization craft, from chart design through dashboards and charting libraries
• Federal information technology experience
• Familiarity with AI-assisted developer tooling

Benefits

Medical, dental, and vision with the employee premium fully paid and half of dependent premiums; employer-paid life, accidental death, and short-term and long-term disability insurance; a 401(k) matched 100% up to 4% of salary, vesting immediately; unlimited paid time off; and sponsored professional certifications and continuing education.