`

Preparation for the Data Scientist role at Argus Media requires a balanced approach that covers deep technical skills, domain curiosity, and structured problem-solving. Review the key evaluation criteria below to align your preparation with what the hiring team values most.

Role-Related Knowledge – You must demonstrate a strong command of statistical modeling, machine learning algorithms, and mathematical optimization. Expect to show deep familiarity with R or Python, as well as the ability to write clean, modular, and reproducible code under time constraints.

Mathematical & Quantitative Rigor – The team values candidates who understand the "why" behind the algorithms. You should be prepared to explain the mathematical foundations of your models, discuss optimization constraints, and explain how you validate predictions in highly volatile markets.

Domain Awareness & Product Curiosity – You do not need to be an energy market expert on day one, but you must show a strong interest in commodity markets. Take the time to understand how Argus Media operates as a price reporting agency and how data science enhances their market analysis.

Resilience & Communication – The interview process can involve ambiguous requirements and intensive technical tasks. You will be evaluated on how clearly you explain your technical decisions, how you handle feedback on your code, and your ability to collaborate across quantitative and non-technical teams.

The hiring process for a Data Scientist at Argus Media is rigorous and highly focused on practical technical capability. The process typically begins with an initial HR screening call to discuss your background, location preferences, and work authorization. This is followed by a heavy emphasis on take-home technical assessments, which are designed to evaluate your hands-on coding and modeling skills before you meet the broader team.

Once you pass the initial technical screening, you will progress to virtual interviews with the hiring manager and senior members of the data science team. These rounds involve a mix of live coding, theoretical discussions, and deep dives into your previous projects. The pace of the process can vary significantly depending on the location and specific team, with some candidates experiencing rapid turnarounds and others navigating a more extended timeline.

`

`

This visual timeline represents the standard path a candidate takes from application to offer. Use this overview to budget your preparation time, ensuring you allocate sufficient energy to the intensive take-home assessment stage, which acts as the primary gateway to the technical panel interviews.

To stand out during the selection process, you must excel in three core areas that the hiring team evaluates closely.

R Programming & Statistical Computing

At Argus Media, R is a critical language used for quantitative modeling, statistical analysis, and generating forward curves. Your ability to write efficient, clean, and well-documented R code is tested extensively through take-home assignments and live coding sessions.

Be ready to go over:

• Tidyverse and Data Manipulation – Efficiently filtering, grouping, and transforming complex datasets.
• Package Development & Reproducibility – Structuring your code so that other data scientists can easily run and verify your models.
• Performance Optimization – Identifying bottlenecks in your code and utilizing vectorization to speed up computations.
• Advanced concepts (less common) – Integrating R with other languages, building interactive dashboards, and writing custom statistical tests for niche market distributions.

Example scenarios:

• Refactoring a slow, loop-heavy script into an optimized, vectorized process.
• Designing a robust data cleaning pipeline that handles missing values and anomalies in daily transaction logs.

Optimization & Quantitative Modeling

The core value proposition of the data science team lies in creating forward-looking market representations. You will be evaluated on your ability to frame physical market constraints as mathematical optimization problems.

Be ready to go over:

• Linear & Non-Linear Programming – Formulating objective functions and defining realistic physical constraints.
• Time-Series Forecasting – Applying autoregressive models, state-space models, or modern machine learning approaches to sequence data.
• Risk & Volatility Modeling – Understanding how price spreads, seasonality, and geopolitical factors impact market curves.
• Advanced concepts (less common) – Stochastic programming, multi-objective optimization, and dynamic hedging formulations.

Example scenarios:

• Explaining how you would construct a forward curve for a commodity with highly seasonal demand and limited storage capacity.
• Developing an optimization model to minimize transportation costs across a multi-node shipping network.

Machine Learning Theory & Modern Methodologies

Beyond classical statistics, the team leverage modern machine learning techniques to extract insights from unstructured data. You will face rigorous theoretical questioning to ensure you understand the mechanics of the models you deploy.

Be ready to go over:

• Model Evaluation Metrics – Choosing the right loss functions and validation strategies for imbalanced or highly skewed data.
• Natural Language Processing (NLP) – Utilizing language models to parse market reports, news feeds, and regulatory filings.
• LLM Evaluation – Understanding techniques like LLM as a judge to automate the quality control of generated summaries or data extractions.
• Advanced concepts (less common) – Deep learning architectures for time-series, transfer learning in low-data regimes, and custom loss function design.

Example scenarios:

• Discussing how you would design an evaluation framework to test if an LLM is accurately extracting commodity prices from news articles.
• Explaining how to prevent data leakage when training a machine learning model on sequential, time-dependent market data.

`

`

As a Data Scientist at Argus Media, your day-to-day work will be highly collaborative, analytical, and aligned with core business outputs. You will be responsible for:

• Developing and Maintaining Forward Curves – Building, testing, and refining mathematical models that generate Possibility Curves and other predictive market indicators.
• Collaborating with Market Analysts – Working closely with domain experts, editors, and economists to translate qualitative market insights into quantitative model parameters.
• Building End-to-End Data Pipelines – Designing scalable pipelines to ingest, clean, and process structured and unstructured market data from global sources.
• Deploying Production-Grade Models – Ensuring that your machine learning and optimization models are robust, well-tested, and integrated seamlessly into production systems.
• Presenting Insights to Stakeholders – Communicating complex technical findings, model limitations, and performance metrics clearly to both internal product teams and external clients.

To be competitive for this position, you should possess a strong quantitative background coupled with practical software engineering sensibilities.

• Must-have skills – Proficient in R or Python with a strong emphasis on statistical computing, experience in mathematical optimization, solid grounding in classical machine learning, and experience with time-series forecasting.
• Nice-to-have skills – Prior experience working in energy, commodity, or financial markets; familiarity with NLP and Large Language Models (LLMs); experience with cloud platforms (AWS, Azure, or GCP); and knowledge of version control and CI/CD pipelines.
• Experience level – Typically requires a Master's or PhD in a quantitative discipline (such as Statistics, Mathematics, Operations Research, Physics, or Quantitative Finance) or equivalent practical industry experience.

`