Senior Quantitative Sports Modeler

Underdog; United States; On-site; USD 135,000 - 185,000; Full time

Benefits offered by this job

Unlimited PTO

Parental leave

Home office stipend

Virtual-first culture

Health plan

Job summary

Underdog is seeking a seasoned Sports Modeler to join our team as a senior contributor to the quantitative engine behind our prediction market operations. You’ll own high-impact models from first research to production monitoring, collaborating with Trading and Engineering to production‑ize ideas with rigor.

You will apply Bayesian methods to synthesize internal estimates with external prices and trader observations, ensuring robust fair-value views while preserving data integrity and

Qualifications

• Front-runner in sports modeling and quantitative research.
• Proven ability to build calibrated probabilistic models for markets.
• Strong knowledge of Bayesian methods and backtesting.

Responsibilities

• Design, build, and validate simulation-based predictive and pricing models across sports and markets.
• Translate sport mechanics and data into outcome distributions and probabilities.
• Own model lifecycle from framing to deployment and monitoring.
• Apply Bayesian inference to update priors with new information and assess price changes.

Skills

Python

Bayesian inference

GitHub

Sports modeling

Market dynamics

Code debugging

Team collaboration

AI tools

Tools

GitHub

Job description

At Underdog, we make sports more fun.

Our thesis is simple: build the best product for sports fans and we will win the space. Founded in 2020, we're one of the fastest-growing sports companies ever, and there’s still so much left to do. We've built and scaled products across fantasy sports, sports betting, and prediction markets, all united in one app that’s seamless, intuitive, and actually fun to use.

Underdog isn’t for everyone. One of our core values is give a sh*t. The people who win here are the ones who care, push, and perform. We’re remote-first by design, but highly collaborative with regular in-person gatherings that help teams build, iterate, and move fast together.

Got that dog in you? Join us.

After all, winning as an Underdog is more fun.

What you'll do:

As a Sports Modeler at Longshot Capital, you will be a senior individual contributor on the team responsible for the quantitative engine behind our prediction market trading operation. Reporting to the Senior Manager, Pricing, you will own high-impact models from initial research through production monitoring, while helping set the technical standard for the wider modeling group. Your work will be tested in live markets and evaluated through calibration, market quality, and trading performance.

This role is for someone who combines deep sports modeling expertise with strong market judgment. You will use Bayesian methods to synthesize internal model estimates, sportsbook prices, exchange market information, and trader observations into the best available view of fair value. You will work hands‑on with Trading and Engineering, using AI tools aggressively and responsibly to shorten the path from idea to production without lowering the bar on rigor or correctness.

Build Simulation Based Sports Models

• Design, build, and validate simulation‑based predictive and pricing models across multiple sports, market types, and time horizons, including pregame and live opportunities where appropriate.
• Translate sport mechanics, team and player data, correlations, and uncertainty into full outcome distributions and market probabilities rather than relying only on point estimates.
• Own the model lifecycle from problem framing, data and feature design through backtesting, calibration, deployment, documentation, and ongoing production monitoring.
• Apply Bayesian inference to establish and update priors as new information arrives, with a clear understanding of when the evidence supports a meaningful change in price.

Combine Model Outputs with Market Information

• Understand how sportsbook prices are formed, including margin, market limits, timing, line movement, and differences in information quality across operators.
• Interpret exchange prices and order books in the context of liquidity, spread, depth, flow, adverse selection, inventory, and other market‑microstructure effects.
• Combine internal model outputs, sportsbook signals, exchange information, and structured trader feedback into defensible estimates of fair value, while quantifying uncertainty and avoiding false precision.
• Measure performance against market benchmarks and trading outcomes, diagnose systematic weaknesses, and prioritize improvements with the greatest expected commercial impact.

Use AI to Increase Development Velocity

• Use AI‑assisted workflows throughout research, prototyping, coding, testing, debugging, review, and documentation to move quickly from a modeling idea to a production‑ready solution.
• Give AI tools precise technical context, constraints, and acceptance criteria; critically evaluate their output; and retain full ownership of methodology, code quality, and correctness.
• Write clear, maintainable Python and use Git and GitHub effectively for branching, pull requests, code review, version control, testing, and collaborative development.
• Build reusable modeling components and improve runtime, reliability, and reproducibility so the team can iterate quickly without accumulating avoidable technical debt.

Partner Across Trading Engineering and Product

• Build a tight feedback loop with traders, turning real‑time market observations and recurring pricing issues into testable hypotheses and concrete model improvements.
• Work with engineers to move models into production reliably, define data and infrastructure requirements, and resolve performance or operational issues when they arise.
• Support new sports, markets, and product launches by assessing model feasibility, identifying key sources of uncertainty, and delivering pricing solutions on practical timelines.
• Explain complex modeling choices, limitations, and tradeoffs clearly to technical and non‑technical partners, and incorporate constructive challenge into the work.

Provide High Level Technical Leadership

• Raise the technical bar through model and code reviews, clear design documentation, thoughtful challenge, and hands‑on support for other modelers.
• Help define modeling standards and reusable approaches across sports while recognizing when a sport or market requires a purpose‑built solution.
• Operate effectively when priorities change, information is incomplete, and timelines are compressed; make sound decisions, communicate risks, and keep work moving.

Who you are:

• At least 5 years of experience in sports modeling, quantitative research, predictive analytics, or a closely related field, including substantial ownership of production models.
• A demonstrated record of building simulation‑based sports models that produce calibrated probabilities and can be evaluated against real markets or real‑money decisions.
• Deep knowledge of probability, statistical modeling, Bayesian inference, simulation, backtesting, calibration, and model validation.
• Strong understanding of both sportsbook market dynamics and exchange‑market dynamics, and the judgment to combine market information with independent model estimates.
• High proficiency in Python and practical experience using Git and GitHub in a collaborative production‑code environment.
• An AI‑forward mindset and advanced, demonstrable use of AI coding and research tools, including the technical depth to direct them precisely and verify their work rather than accept output at face value.
• Strong ownership, attention to detail, and the ability to thrive in a fast‑paced environment with frequent change, incomplete information, and competing priorities.
• Clear communication and strong cross‑functional instincts, with the ability to work effectively with modelers, traders, engineers, product partners, and business leaders.
• Willingness to work occasional evenings and weekends during major sports periods and other time‑sensitive launches, particularly around the NFL season.

Bonus points if you have:

• Direct experience building models for a market maker, proprietary trading firm, prediction‑market business, exchange, or sportsbook trading desk.
• Experience trading on exchanges or building pricing, liquidity, or automated market‑making solutions for exchange venues.
• Experience with in‑play or other low‑latency models where real‑time state, reliability, and execution speed materially affect performance.
• Experience designing modeling infrastructure, cloud‑compute solutions, backtesting systems, production monitoring, or GitHub‑based CI/CD pipelines and workflows.
• Experience optimizing computationally intensive code or building performance‑critical components in C++ or Rust.
• A genuine passion for sports and curiosity about how information moves prices.

*Our target starting base salary range for this position is between $135,000 and $185,000, plus target equity.*Final compensation will be determined based on a variety of factors including relevant experience, skills and expertise, interview performance, and other market considerations.

What we can offer you:

• Unlimited PTO for full‑time employees (we're extremely flexible with the exception of the first few weeks before and into the NFL season)
• 16 weeks of fully paid parental leave
• Home office stipend
• A virtual‑first culture, but we come together a few times per year in person for offsites, trainings, hackathons, and more
• Slack reactions for every occasion
• 401k with up to 5% employer match
• Company paid health, dental, vision plan options for employees and dependents, FSA, and more
• Opportunity to celebrate your favorite sports teams through local meetups and company‑sponsored events
• Claude and ChatGPT enterprise accounts

#LI-REMOTE

Underdog is an equal opportunity employer and doesn't discriminate on the basis of creed, race, sexual orientation, gender, age, disability status, or any other defining characteristic.

California Applicants: Review our CPRA Privacy Notice here.