The ad says“Experience designing and analyzing experiments (A/B testing) and applying statistical inference or causal methods”
The ad says“Experience with LLM evaluation and observability tools like Langfuse, Braintrust, or internal harnesses, and with building automated evaluators”
The ad says“Experience designing and analyzing experiments (A/B testing) and applying statistical inference or causal methods”
The ad says“Experience with modern data stack tools like dbt or equivalent”
The ad says“Experience with notebook tools like Jupyter, Hex, Hyperquery, or equivalent”
The ad says“Experience with notebook tools like Jupyter, Hex, Hyperquery, or equivalent”
The ad says“Experience with notebook tools like Jupyter, Hex, Hyperquery, or equivalent”
The ad says“Experience with LLM evaluation and observability tools like Langfuse, Braintrust, or internal harnesses, and with building automated evaluators”
The ad says“Experience fine-tuning, distilling, or rigorously evaluating LLMs, or applying causal inference and experimentation at scale”
The ad says“Experience with LLM evaluation and observability tools like Langfuse, Braintrust, or internal harnesses, and with building automated evaluators”
The ad says“Experience fine-tuning, distilling, or rigorously evaluating LLMs, or applying causal inference and experimentation at scale”
The ad says“Experience with LLM evaluation and observability tools like Langfuse, Braintrust, or internal harnesses, and with building automated evaluators”
The ad says“Strong proficiency in Python and common ML and statistics libraries (e.g., scikit-learn, PyTorch, pandas, statsmodels)”
The ad says“Strong proficiency in Python and common ML and statistics libraries (e.g., scikit-learn, PyTorch, pandas, statsmodels)”
The ad says“Strong proficiency in Python and common ML and statistics libraries (e.g., scikit-learn, PyTorch, pandas, statsmodels)”
The ad says“Strong proficiency in Python and common ML and statistics libraries (e.g., scikit-learn, PyTorch, pandas, statsmodels)”
The ad says“Strong proficiency in SQL”
The ad says“Experience designing and analyzing experiments (A/B testing) and applying statistical inference or causal methods”
The ad says“Strong proficiency in Python and common ML and statistics libraries (e.g., scikit-learn, PyTorch, pandas, statsmodels)”
Traba is hiring a Senior Data Scientist to join the founding Agents team and lead measurement and modeling for our agentic platform from 0→1
You’ll make the core calls on how agent quality is defined, measured, and improved; set the bar for statistical and scientific rigor; and build the evaluation, experimentation, and modeling foundations that every agent we ship is measured against
As a Senior Data Scientist at Traba, you’ll own how we model and understand agent performance inside real customer workflows—capability, reliability, and unit economics—and partner with engineering, product, and operations leadership on the decisions that shape the platform
Provide strategic insights and recommendations to senior leadership through in-depth statistical analysis and modeling
Design, build, and maintain the metrics, models, and reporting that track agent quality, reliability, adoption, and unit economics for stakeholders across the Agents and Operations teams
Build evaluation and experimentation as a first-class discipline—datasets from production traces, rubrics, automated graders, regression suites, and the experiment design and analysis that prove causation—so every agent improvement we ship is backed by evidence
Identify key business challenges and opportunities—including agent failure modes, tool-use patterns, and cost and latency—and build statistical and machine-learning models to drive product improvements and growth initiatives
Architect scalable analytics and modeling infrastructure to ensure data integrity, governance, and accessibility for both human and agent consumers
Oversee the development and maintenance of Traba’s data warehouse to ensure data availability and governance
Work closely with the Agents team and Operations leadership to understand their data needs and provide actionable, statistically grounded insights that drive continuous process improvement and operational efficiency
Provide Operations teams with models, self-service analytics, and advanced technologies—including AI-assisted tools—enabling them to independently analyze operational data and optimize their daily activities
Mentor the scientists and analysts who build alongside you, and set the standards that define what “good” looks like for measurement, modeling, and experimentation at Traba
Benefits
Competitive salary
Early stage start-up equity
100% paid health, dental & vision coverage
Dinner Provided via DoorDash, free DashPass & stocked kitchen for NY employees
One Medical Membership
Flexible PTO
Parental leave
Commuter benefit
HSA via Optum
Mental health benefits: Talkspace, HealthAdvocate, Teledoc Health
Wellhub (Gympass) Membership
Preferred experience
Strong proficiency in SQL
Curiosity and Initiative: Intense curiosity to ask “why?” and use data to find answers, combined with a “no task too small” mentality
Strong proficiency in Python and common ML and statistics libraries (e.g., scikit-learn, PyTorch, pandas, statsmodels)
Education: BS/MS/PhD in data science, statistics, machine learning, computer science, mathematics, economics, or a related quantitative field (or equivalent work experience)
Communication Skills: Excellent data storytelling skills to effectively engage with stakeholders
Technical Skills:
Self-Motivation: Ability to work independently and as part of a team in a fast-paced startup environment
Experience designing and analyzing experiments (A/B testing) and applying statistical inference or causal methods
Experience: 4-8 years in data science, machine learning, applied statistics, or quantitative research, with 2+ years of hands-on work modeling or measuring LLM- or agent-based systems in production
Collaboration Skills: Strong ability to work across departments, identifying and prioritizing analytics problems to deliver actionable insights
Experience with LLM evaluation and observability tools like Langfuse, Braintrust, or internal harnesses, and with building automated evaluators
Experience with notebook tools like Jupyter, Hex, Hyperquery, or equivalent
Experience with modern data stack tools like dbt or equivalent
Experience building internal agents or MCP servers for analytics workflows, or prior work at a vertical AI or AI-native data company (e.g., Hex, Omni, dbt)
Experience fine-tuning, distilling, or rigorously evaluating LLMs, or applying causal inference and experimentation at scale
