Skip to main content

Personas

There are a variety of people which may interact with Nebari deployments. Below is a list of fabricated personas intended to cover most people who interact with Nebari.

User groupsโ€‹

Each persona is placed in a Nebari user group:

End users are the folks who are doing their daily work on Nebari. They are using Nebari as a platform to complete a variety of tasks from software development to dashboard creation to machine learning training and evaluation. Our end users are Alia, Noor, Robin, and Skyler.

Super users, like end users may also be using Nebari for their daily work, but they do so with elevated privileges. These folks are in charge of managing teams of people and as such may be managing environments for their team. Our super users are Blake, Enzo, Jacob, and Sam.

Customer represents the person at an organization making the high level decision to adopt Nebari as a platform for their employees. Our customer is Jordan.

Admin represents the people who are in charge of deploying Nebari. They are responsible for maintaining and running the day to day maintenance. This may include cloud account management, conda-store maintenance, kubernetes cluster management and adding/removing end users from the deployment. Our admin is Taylor

Personasโ€‹


Aliaโ€‹

Age bracket: 25-35

Role: Data scientist

Reports to: Head of Data Science

Nebari user group: end user

cartoon image of persona alia

๐Ÿ“ฃ Storyโ€‹

  • As a Data scientist I want to investigate, extract and report meaningful insights with my organisationโ€™s data.
  • As a DS I want to build self-serving internal data products to make data simple within the company.

๐Ÿ”ง Software toolsโ€‹

  • Python - NumPy, Pandas, scikit-learn, etc.
  • Airflow
  • GCP buckets
  • SQL
  • Data viz and dashboarding tools like Plotly, Panel, Altair...
  • VSCode, Jupyter notebooks

๐ŸŒฎ Core needsโ€‹

N/A

๐Ÿ› Pain points or biggest challengesโ€‹

  • Poor data quality and quantity

  • High friction in collaborative projects

  • Ability to experiment with new methods, libraries, and datasets in an efficient and quick way

  • Access to multiple data sources with minimal friction

  • Clear communication channels with software engineering, infrastructure, and other departments

  • Need to communicate results to non-technical people


Blakeโ€‹

Age bracket: 25-35

Role: Software Engineer/ ML Engineer

Reports to: Head of Software Engineering

Nebari user group: super user

cartoon image of persona blake

๐Ÿ“ฃ Storyโ€‹

  • Nebari is a place I go to do my everyday programming work. I use conda envs, jupyter, and write code in vscode. I deploy dashboards with cds dashboards. I create packages and I need to pip install the dev versions into my environment.
  • As a SWE I want to be able to collaborate with the Data and Engineering teams to build new prototypes for products

๐Ÿ”ง Software toolsโ€‹

  • VSCode
  • Jupyter notebooks
  • conda, pip
  • Python and stack
  • Plotly, Altair, etc.

๐Ÿ› Pain points or biggest challengesโ€‹

N/A

๐ŸŒฎ Core needsโ€‹

  • Need to be able to create prototypes efficiently and run experiments with product and data teams
  • ability to ship products without being held up in the process
  • ability to manage my environment and tooling

Enzoโ€‹

Age bracket: 30-40

Role: Staff Machine learning engineer at a startup

Reports to: Head of Software Engineering

Nebari user group: super user

cartoon image of persona enzo

๐Ÿ“ฃ Storyโ€‹

  • As an ML Engineer in a startup, I have good modeling and coding skills, but I want to build MLOps pipelines as easily as possible, without having to learn DevOps first, so that I can efficiently test and productionize my models.

๐Ÿ”ง Software toolsโ€‹

  • Python, PyTorch
  • Git, GitHub
  • AWS, GCP
  • Argo
  • DVC, ClearML, Tensorboard
  • CI/CD platform
  • K8s, Dask

๐Ÿ› Pain points or biggest challengesโ€‹

  • Lack of DevOps experience makes pipeline implementation difficult, time consuming, and error prone.
  • Limited high-level access to pipeline tools means that they are essentially black boxes to me (canโ€™t see logs, not sure whatโ€™s happening when something breaks)

๐ŸŒฎ Core needsโ€‹

  • Safe and efficient access to infrastructure without breaking things
  • Elevated privileges to experiment with different configurations and for transparency into background tasks
  • Documentation that shows how to implement common pipeline configurations
  • Templates/base configurations with several different scenarios

Jacobโ€‹

Age bracket: 30-40

Role: Staff Machine learning engineer

Reports to: Head of Software Engineering

Nebari user group: super user

cartoon image of persona jacob

๐Ÿ“ฃ Storyโ€‹

  • As an ML Engineer with good DevOps skills and not much time, I want to build MLOps pipelines efficiently so that I donโ€™t waste time on tedious tasks.

๐Ÿ”ง Software toolsโ€‹

  • Python, PyTorch
  • Git, GitHub
  • AWS, GCP
  • Argo
  • DVC, ClearML, Tensorboard
  • CI/CD platform
  • K8s, Dask

๐Ÿ› Pain points or biggest challengesโ€‹

  • Rigid pipelining tools limit flexibility to build pipelines that meet my specific constraints.
  • Limited or incomplete documentation (just hello world examples) doesnโ€™t give me enough information to build the complex configuration I need without a lot of trial and error testing.

๐ŸŒฎ Core needsโ€‹

  • Efficient access to infrastructure
  • Elevated privileges to experiment with different configurations and for transparency into background tasks
  • Documentation that shows how to implement common pipeline configurations
  • Templates/base configurations with several different scenarios
  • Flexible pipelining platform that allows experimentation and custom solutions

Jordanโ€‹

Age bracket: 45-60

Role: CEO

Reports to: โ€”

Nebari user group: customer

cartoon image of persona jordan

๐Ÿ“ฃ Storyโ€‹

  • As a CEO I am interested in knowing how the data teams are helping us improve our processes and revenue streams.
  • I donโ€™t know anything about programming or jupyter. As the CEO I want to click on a button or email, sign in, and see informative dashboards and metrics.
  • I want to have a clear breakdown of costs from my departments.

๐Ÿ”ง Software toolsโ€‹

  • Spreadsheets
  • Email
  • Data dashboards

๐Ÿ› Pain points or biggest challengesโ€‹

  • there is big churn in the industry right now - how can we easily bring up new hires to speed?

๐ŸŒฎ Core needsโ€‹

  • low-barrier access to data insights and outputs
  • easy to track budgets and costs
  • please stop the fights on data and swe departments

Noorโ€‹

Age bracket: 16-25

Role: Student Python Programmer

Reports to: โ€”

Nebari user group: end user

cartoon image of persona noor

๐Ÿ“ฃ Storyโ€‹

  • As a student, I want to learn how to use Python to get a job in software development after I graduate.
  • Iโ€™m learning how to code. I know I need an environment, but someone else creates and manages them. I never modify an environment myself (either through conda or pip)

๐Ÿ”ง Software toolsโ€‹

  • Python
  • VScode and Jupyter notebooks
  • conda, pip

๐ŸŒฎ Core needsโ€‹

  • simple and intuitive platform
  • easy sharing and collaboration
  • ability to experiment without worrying about โ€œbreaking thingsโ€

๐Ÿ› Pain points or biggest challengesโ€‹

  • I do not know how to share my work with friends and peers

Robinโ€‹

Age bracket: 45-55

Role: Head of Data Science

Reports to: CTO

Nebari user group: end user

cartoon image of persona robin

๐Ÿ“ฃ Storyโ€‹

  • As the head of Data Science I want to make sure my team has all the tools and data needed to be successful in their job.
  • As the HDS I want to have executive buy-in to grow the data practice in house to improve our internal processes and better serve our customers.
  • As the HDS I want to not have to spend half of my budget in proprietary platforms.
  • As the HDS I want to work with organisation leadership to help establish long-term vision, strategy and roadmap for one or more teams

๐Ÿ”ง Software toolsโ€‹

  • Git/ GitHub
  • R and Python
  • Dhasboarding and data viz tools
  • Spark, SQL
  • VSCode, Jupyter notebooks

๐ŸŒฎ Core needsโ€‹

  • Low friction tooling to make my team more productive
  • Foster collaboration, mentorship and peer-programming
  • A platform that โ€œjust worksโ€
  • An easy way to share insights with non-technical executives and folks at the company

๐Ÿ› Pain points or biggest challengesโ€‹

  • Lack of alignment across departments and lack of OKRs
  • Often get caught on discussions between my team, software engineering, and infrastructure to move items to production
  • Another resignation - need to offboard and onboard people seamlessly

Samโ€‹

Age bracket: 30-40

Role: Research Data Scientist

Reports to: Head of Data Science

Nebari user group: super user

cartoon image of persona sam

๐Ÿ“ฃ Storyโ€‹

  • As a research data scientist I want to use ML to measure and optimise the costs, performance, efficiency and reliability of our companyโ€™s infrastructure to deliver the best experience to our customers.
  • As a research data scientist I want to implement and test out new approaches both on toy test tasks as well as on actual application scenarios.

๐Ÿ”ง Software toolsโ€‹

  • Python
  • Machine Learning frameworks
  • SQL
  • Dask
  • VSCode, Jupyter notebooks

๐ŸŒฎ Core needsโ€‹

  • Collaboration with engineering, product and the rest of the DS team
  • Ability to track experiments, models, data, accuracy, metrics reliably
  • Ability to access scalable compute on-demand

๐Ÿ› Pain points or biggest challengesโ€‹

  • Cannot scale the compute resources without having to go through infrastructure approval
  • Not easy way to track experiments, data lineage, and workflow artefacts
  • Do not understand git well enough

Skylerโ€‹

Age bracket: 35-45

Role: Staff Machine learning engineer

Reports to: Head of Software Engineering

Nebari user group: end user

cartoon image of persona skyler

๐Ÿ“ฃ Storyโ€‹

  • As a ML engineer I want to work across teams and functions -- software engineers, ML engineers, data scientists, researchers, and product managers to understand user problems and build engineering solutions.
  • As a ML Engineer I want to work with ML engineers and related teams to integrate ML components into the system and solve their performance issues

๐Ÿ”ง Software toolsโ€‹

  • Docker
  • CI/CD platforms
  • Python, Go, bash
  • VSCode and Vim
  • Flyte, Kubeflow
  • GCP, Azure, AWS, DO
  • Dask
  • Git, GitHub, GitLab

๐ŸŒฎ Core needsโ€‹

  • highly collaborative and extensible tooling and platform
  • enforcement of best practices
  • safe and efficient access to both the data and the infrastructure
  • elevated privileges to experiment in โ€œstaging or pre-productionโ€ environments

๐Ÿ› Pain points or biggest challengesโ€‹

  • Lack of best practices across other sides of the organisation makes my job really hard
  • I am not sure if the models in production have bugs or the data is no longer accurate, need better monitoring tools

Taylorโ€‹

Age bracket: 30-45

Role: Sr. Dev Ops Engineer

Reports to: Head of Software Engineering

Nebari user group: admin

cartoon image of persona taylor

๐Ÿ“ฃ Storyโ€‹

  • I maintain my orgโ€™s Nebari deployment. I make sure the infrastructure as code is maintained and that everyone has the conda environments or other resources they need.
  • As a DevOps engineer I want to automate away as much of the day to day as possible - โ€œRun By Robotsโ€ is the goal to minimise issues and make out team more efficient.
  • Identify, respond to and collaborate with support and product teams to resolve production and customer issues and incidents.

๐Ÿ”ง Software toolsโ€‹

  • GCP, Azure, AWS, DO
  • GNU/Linux
  • Scripting - bash
  • Terraform, Ansible, Docker...
  • Kubernetes
  • Grafana, Prometheus, Elastic

๐Ÿ› Pain points or biggest challengesโ€‹

  • The organisation still relies on many manual operations
  • There is so much technical debt - we can lose all of our infrastructure easily

๐ŸŒฎ Core needsโ€‹

  • Rely on automation as much as needed
  • Have a product and security first approach to data and SWE
  • Have organisation wide best practices for monitoring, alerting and incident management
  • Ensure the whole company adheres to best practices