Data Scientist IC2 works confidently with data and models. Learns quickly and progresses consistently while requiring minimal feedback from senior data scientists. Develops the basics across all three of the technical areas. Starts to understand the structure of the tribe/squad is working in and establishes connections with other data scientists and engineers and PMs in other teams of the same squad/tribe.

65,000-85,000 GBP

50,000-65,000 EUR (Estonia)

145,000-185,000 USD (in New York)

Senior Data Scientist identifies high-impact ML opportunities across the company, works effectively with multiple domains and aims for cross-company impact. Leads team through making well-reasoned decisions. Develops and defends an approach while considering pros and cons of multiple implementations. Applies knowledge of industry trends and best practices. Develops deeper expertise across all three of the technical areas.

85,000-115,000 GBP

65,000-83,000 EUR (Estonia)

185,000-230,000 USD (in New York)

Senior +: Recognized at tribe level as specialists in at least one area of Data Science, Staff Data Scientists must drive projects that: - solve bottlenecks for DS productivity in their squad(s) or go in-depth within one domain - increase scientific rigour, observability and governance - introduce new DS use cases - improve ease of work for DS in their squad(s) and mentor more junior team members to increase their quality of work They can mentor 1-3 peers and lead 1-2 peers, and are expected to conduct training for a wider audience. They work closely with DS Leads and Senior Leads to create the vision and roadmap for Data Science in their squad(s) and tribe.

115,000-150,000 GBP

83,000-105,000 EUR (Estonia)

230,000-280,000 USD (in New York)

Data Science Lead 1 effectively guides a small team of data scientists focused on a specific product or feature. They facilitate project planning, execution, and delivery, ensuring alignment with product goals. They foster a collaborative environment, provide technical mentorship, and champion effective communication within the team and with stakeholders.

85,000-115,000 GBP

65,000-83,000 EUR (Estonia)

185,000-230,000 USD (in New York)

Senior +: Requires organizational and interpersonal skills. Lead Data Scientist focuses on growing other data scientists and working with them to solve cross-company data science problems. Coaches multiple data scientists and supports them in their progress. Thinks long-term and helps the team to scale. Discusses statistical, software engineering and data issues with reports. Mentors, pair programs, and conducts code reviews. Managerial focussed DS also directly leads the work of a team, generally focused on a tribe or group function. Their effectiveness is measured in the output and happiness of their team, not just their own individual contribution.

115,000-150,000 GBP

83,000-105,000 EUR (Estonia)

230,000-280,000 USD (in New York)

Lead +: Sets strategic direction for the DS discipline. Active participant in leadership of the entire data group. Communicates consistently accross the company. Can coomunicate with DS coutnerparts at external paterns and vendors. Coordinates with all of the product managers accross the company to make sure that their DS needs are being met and to proactively define and build new DS products. Actively promotes the company externally, from recruiting, culture, customer, data, and technology perspectives.

150,000-175,000 GBP

105,000-124,000 EUR (Estonia)

280,000-325,000 USD (in New York)

IC1: Works comfortably with SQL - querying, aggregating, transforming and operating with data. Learns and uses BI tools like LookML, Superset.

IC2: Advanced use of a SQL - window functions, data-type based (arrays, JSON, etc), creation of custom functions. Learns to access and process data on a variety of databases and parallel data processing technologies.

IC3: Understands the inner workings of various database and parallel processing technologies and their pros and cons. Conscious of infrastructure costs when building and executing tasks. Capable of doing ETL on data with different formats and types. Has an end to end understanding and vision of the data model.

IC4: Understands full lifecycle of data in their squad(s) and is able to propose and carry out improvements to its architecture. Has previous experience building large-scale streaming and batch pipelines.

DSL1:Understands the inner workings of various database and parallel processing technologies and their pros and cons. Conscious of infrastructure costs when building and executing tasks. Capable of doing ETL on data with different formats and types. Has an end to end understanding and vision of the data model.

DCL2: Generalist. Understands multiple data domains in the company. Serves as an acknowledged leader for understanding the inner workings and the debugging of complex queries and workflows on paralell data processing systems.

DCL3: Generalist. Understands multiple data domains in the company. Serves as an acknowledged leader for understanding the inner workings and the debugging of complex queries and workflows on paralell data processing systems.

Software Engineering

IC1: Knows how to work with atleast one high-level programming language, and comfortably works with data science specific libraries in Python/R. Familiar with using git for both submitting and reviewing PRs; practice clean coding and writing docs (with support from their leads)

IC2: Uses code to automate part of his/her work or analysis. Masters tools like version control, Docker, CI/CD pipelines. Works autonomously and is capable of unblocking themselves by acquiring knowledge through documentation.

IC3: Understands software engineering principles. Understands how different components and pieces fit into the architecture. Writes clean, well-formatted, and well-structured code. Understands how to organize and develop a new software package. Can deploy custom APIs. Is aware of the overall architecture and can suggest improvements.

IC4: Specialist in at least Python and SQL and knows a few other languages. Builds high quality software libraries that are used across the tribe and company. Understands how to make decisions about the best paradigm/design for the problem at hand, e.g. OOP vs Procedural. Makes large impact contributions and thorough code review for peers. Keeping in mind quality and ease of maintenance moving forward.

DCL1:  Understands software engineering principles. Understands how different components and pieces fit into the architecture. Writes clean, well-formatted, and well-structured code. Understands how to organize and develop a new software package. Can deploy custom APIs. Is aware of the overall architecture and can suggest improvements.

DCL2: Leads the development of complex codebases worked on by multiple DS and/or engineers. Guides team philosophy on software engineering best practices.  

DCL3: Leads the development of complex codebases worked on by multiple DS and/or engineers. Guides team philosophy on software engineering best practices.

Modelling

IC1: Has developed simple models using standard ML libraries. Comfortable with validation side of things. Understands overfitting, training/testing, data leaks.

IC2: Can develop new models or improve existing ones to help the team achieve its goals. Automates model processes and enhances model governance e.g. feature development, retraining pipelines, parameter tuning, testing, scheduling, scoring, output summarization.

IC3: Develops deeper knowledge of multiple ML techniques, their strengths and weaknesses, and when each technique is appropriate. Reads and understands statistics and ML reserach papers and tries to apply them to everyday tasks and projects. Has implemented ML techniques previously untested in the company and/or has applied ML techniques to areas where they were not previously used.

IC4: Expert in at least one ML paradigm (e.g. Deep Learning) relevant for their squad(s). Capable of applying novel modelling methods that are not available in open source. Has very broad knowledge in adjacent fields. Extensive experience in model lifecycle, from exploration to impact analysis. Understands the limitations imposed by the AI and privacy regulations and proposes solutions.

DCL1:  Develops deeper knowledge of multiple ML techniques, their strengths and weaknesses, and when each technique is appropriate. Reads and understands statistics and ML reserach papers and tries to apply them to everyday tasks and projects. Has implemented ML techniques previously untested in the company and/or has applied ML techniques to areas where they were not previously used.

DCL2: Develops deeper knowledge of multiple ML techniques, their strengths and weaknesses, and when each technique is appropriate. Reads and understands statistics and ML reserach papers and tries to apply them to everyday tasks and projects. Has implemented ML techniques previously untested in the company and/or has applied ML techniques to areas where they were not previously used.

DCL3: Develops deeper knowledge of multiple ML techniques, their strengths and weaknesses, and when each technique is appropriate. Reads and understands statistics and ML reserach papers and tries to apply them to everyday tasks and projects. Has implemented ML techniques previously untested in the company and/or has applied ML techniques to areas where they were not previously used.

IC1: Has basic statistics knowledge: mean, standard deviation, law of large numbers, normal distribution. Can do basic operations such as calculate standard deviation of an average of two random variables.

IC2: Understands the uncertainty inherent in observations and model predictions. Can calculate measures of uncertainty, such as confidence intervals, for any observed quantity. Can interpret measures of model uncertainty where provided by the model.

IC3: Has strong understanding of some of the most common probabilistic concepts: conditional probability, Bayesian thinking, statistical tests, etc.. Has used this in their daily work and has independently and quickly filled knowledge gaps when needed. Can construct uncertainty measures for models. Can build models to predict probability distributions rather than point values, where appropriate.

IC4: Competent in advanced probabilistic modelling methods and maths. E.g. Causal inference, Probabilistic Graphical models, Markov processes are not the things you slightly remember from university, but you applied them in practice.

DSL1: Has strong understanding of some of the most common probabilistic concepts: conditional probability, Bayesian thinking, statistical tests, etc.. Has used this in their daily work and has independently and quickly filled knowledge gaps when needed. Can construct uncertainty measures for models. Can build models to predict probability distributions rather than point values, where appropriate.

DCL2: Has strong understanding of some of the most common probabilistic concepts: conditional probability, Bayesian thinking, statistical tests, etc.. Has used this in their daily work and has independently and quickly filled knowledge gaps when needed. Can construct uncertainty measures for models. Can build models to predict probability distributions rather than point values, where appropriate.

DCL3: Has strong understanding of some of the most common probabilistic concepts: conditional probability, Bayesian thinking, statistical tests, etc.. Has used this in their daily work and has independently and quickly filled knowledge gaps when needed. Can construct uncertainty measures for models. Can build models to predict probability distributions rather than point values, where appropriate.

Scientific Method

IC1: Has a good grasp of the fundamental principles of research, including the ability to formulate research questions, hypotheses and objectives. Uses statistical analysis for the research task at hand and has knowledge of experimental design principles, A/B testing and control groups.

IC2: Can autonomously design experiments for projects in the specific product area. Can decide the best methodology to collect data and apply the necessary pre-processing steps. Designs well-controlled A/B tests, can set the necessary expectations for the length of the test, can interpret test results with clear connections to the product area.

IC3: Has mastered the full research cycle from formulation of the reasearch question/hypotheses all the way to interpretation of experimentation results. Delegates specific tasks of the reserch piece maintaing overall ownership. Acts as advisor on experimental design for data scientists in other teams. Helps with interpretation of experimental results in areas outside the direct product area of interest.

IC4: Reviews, identifies gaps and proposes improvements to methodological choices by their peers. Develops methodological frameworks and tools (e.g. automated power analysis for experiments) to mitigate knowledge gaps.

DCL1:  Has mastered the full research cycle from formulation of the reasearch question/hypotheses all the way to interpretation of experimentation results. Delegates specific tasks of the reserch piece maintaing overall ownership. Acts as advisor on experimental design for data scientists in other teams. Helps with interpretation of experimental results in areas outside the direct product area of interest.

DCL2: Builds the processes for the team to consistently practice the scientific rigour. Sets up peer review practices. Mentors team members to help them set up their experiments. Keeps an eye on the order of experiments in the team to maximise iteration speed.

DCL3: Synchronises research activities, modelling and experimentation between squads. Actively seeks for gaps and blockers to try eliminating them as soon as possible. Mentors leads to help them build better processes in their teams.