Data Scientist

About the role

This position involves developing and deploying data models and products, from initial concept to full production. You will analyze complex datasets to uncover insights that drive new features and build LLM-driven pipelines with robust evaluation methods. The role requires close collaboration with data scientists, engineers, and co-founders to scale solutions for market use.

Key facts

What you'll do

• Construct data models and products for production, including generative models, subscriber behavior predictions, and solutions for TV providers (MVPDs).
• Investigate large, complex datasets to identify trends that lead to new features and contribute to the core Python data science library.
• Conduct research and development for Antenna, exploring new data and methods to address business questions and present findings to senior stakeholders.
• Produce clear, well-structured, testable, and efficient Python code using object-oriented principles, ensuring thorough documentation.
• Troubleshoot distributed systems and enhance code performance and data handling capacity.
• Develop LLM-powered pipelines and agents, documenting failure modes and implemented safeguards.
• Establish evaluation metrics as a primary output, validating model responses, creating clear pass/fail checks, calibrating LLM-as-a-judge rubrics, and using tracing tools to monitor cost, latency, and quality.
• Integrate agentic coding tools into daily tasks, planning, writing tests, and reviewing changes while maintaining independent design, debugging, and defense of your work.
• Partner with various stakeholders, including co-founders, to clearly articulate technical and system design decisions.

Requirements

• At least 2 years of experience developing machine learning models and data products in Python, with the engineering capability to move them from prototype to production.
• Expert proficiency in Python, including strong object-oriented design, software system design, and experience creating high-quality, testable, production-ready code.
• Practical experience with deep learning frameworks (PyTorch or TensorFlow) and a solid understanding of machine learning concepts, the full model development lifecycle, and MLOps principles.
• Hands-on experience with large-scale data processing tools (e.g., Apache Spark/PySpark, Dask) and strong SQL skills for complex datasets.
• Significant experience with cloud platforms (GCP preferred), including deploying, managing, and scaling services (Docker, Cloud Run, GKE) and working with big data systems (Dataproc, BigQuery).
• Excellent problem-solving abilities, skilled in debugging complex distributed systems and optimizing them for performance and scale.
• Advanced English proficiency (B2-C1) with strong communication, teamwork, and consulting skills, capable of explaining complex technical and system design decisions.
• Daily use of agentic coding tools (Claude Code, Cursor, or Codex CLI), focusing on upfront planning, test writing, and instruction setting, with a thorough review of changes. You must also be able to design, debug, and defend your work independently.
• Experience building and deploying LLM-powered agents or pipelines using orchestration frameworks (LangGraph, Pydantic AI, or OpenAI Agents SDK), including custom tool definitions, agent state and memory, and human review steps. You should be able to explain encountered failure modes and implemented guardrails.
• You consider evaluations a key deliverable, validating model responses with structured outputs (Pydantic), building eval sets with clear pass/fail checks, calibrating LLM-as-a-judge rubrics, and using tracing tools (Langfuse, LangSmith, or Braintrust) to track cost, latency, and quality. You can describe an evaluation that identified an issue missed by human review.

Nice to have

• Experience or interest in the Subscription Economy, particularly media and entertainment, or working with media data or data clean rooms.
• Experience with synthetic data generation or advanced generative models (e.g., GANs, VAEs, CTGAN).
• Experience building Python libraries for others or contributing to open-source projects.
• Knowledge of advanced MLOps practices like model monitoring and build automation tools (e.g., Cloud Build, Cloud Run).
• Experience creating custom tool integrations for agents, including tool definitions, routing against internal APIs, input schemas, validation, and safe handling of side effects.
• Experience with advanced evaluation and observability practices such as multi-judge calibration, automated regression suites, and production monitoring of agent quality.
• Familiarity with RAG and context engineering for grounding model responses in proprietary data.
• Experience using LLMs for testing pipelines and QA workflows.

Skills & tools

• Python
• PyTorch
• TensorFlow
• Apache Spark/PySpark
• Dask
• SQL
• GCP
• Docker
• Cloud Run
• GKE
• Dataproc
• BigQuery
• Claude Code
• Cursor
• Codex CLI
• LangGraph
• Pydantic AI
• OpenAI Agents SDK
• Pydantic
• Langfuse
• LangSmith
• Braintrust

Practical notes

This is a remote, mid-level role. Advanced English proficiency (C1) is required for effective communication on technical and system design matters.