AI-Powered Stage
The AI-Powered Stage in the Data Science section is focused on providing applied AI capabilties to the GitLab product.
Data Science
The Data Science section is focused on leveraging ML and AI in the GitLab product and preventing abuse in the application.
Vision
Build a diverse and global development team to support and drive results across the section, while maintaining our values and unique way of working.
Mission
Drive results through iterative development as we add AI, ModelOps and Anti-abuse features into the product. Our teams are data-driven, support dogfooding, and collaboration within GitLab and the wider community.
Stages
Career Development
Giving / receiving feedback
It can be hard to understand how you’re doing in your role, because feedback can come off as formal (annual reviews, 360 surveys, career development conversations, goal check-ins) or casual (in Slack channels, 1-1’s, MR reviews, team meetings.) We receive various kinds of feedback regularly and through different formats, so the type of feedback you’re receiving is not always clear. In order to be more intentional about the types of feedback given, here is a classification chart based on three types of feedback:
| Label | Meaning | Example |
|---|---|---|
| (appreciation) | I want to thank you for doing this, and please do more of it in the future | “I did not expect that you would have created a working group, because you’ve done so, our whole team will benefit from the results.” |
| (coaching) | I’m trying to help you improve a behavior you are already exhibiting or change a behavior that you currently have | “The reports that you give me are very helpful, and in the future we can schedule them for the first of the month to be more consistent.” |
| (evaluation) | Tells you where you stand according to existing standards or expectations | “My expectation was that our decision would be transparent. Since it was not, our team has forgotten the decision, so we must be sure and meet that expectation next time.” |
Engineering Managers
This section lists relevant experience areas for individual contributors interested in the management track, new engineering managers, or existing engineering managers who may be lacking opportunities. This list can be used to identify opportunities in these areas.
Expert hiring manager
Expert hiring manager
- Experience with behavioral interviews
- Screening candidates for your team
- Identifying cultural answers or clarifying vague answers
- Identifying a headcount need in advance
Performance management
Performance management
- Crucial conversations
- Performance improvement plans
- Coaching on improvement areas
- Giving feedback
- Identifying underperformance
Communicating company decisions
Communicating company decisions
- Annual review, calibration sessions, compensation discussions
- Motivating team members on opportunities that come with negatives (borrow requests, engineering allocations, feature change locks)
Product area
Product area
- Triage reports
- Define and monitor productivity metrics, take action if necessary
- Collaborative planning
- Proactively identifying issues or recommending engineering allocations
- Leading an incident in your area
- Proposing and driving a borrow request (reactive)
- Shared OKRs and delivery
Team success
Team success
- Career growth development leading to promotions, mentors, technical interviews, maintainers
- Setting goals based on 360 feedback and career aspirations
- Smooth onboarding process
- Frequent and transparent handbook updates
- Identifying performance indicators for the team
- Becoming a mentor
Achieving consensus
Achieving consensus
- Facilitating a working group
- Participating in stage, sub-department, skip level, and/or engineering manager discussions
- Coordinating the dev on-call
- Experience with being Incident Manager On-Call (IMOC)
- Collaboration with the full product group quad planning
Personal growth
Personal growth
- Receiving feedback
- Continued learning, identifying new growth opportunities, and building a personal growth plan
- Seeking a mentor
Trainings offered by GitLab for EMs
Other resources
- Slack #aspiring-managers channel
- Handbook page dedicated to growing leadership skills for ICs
Staff Engineers
For an explanation on what to expect as a Staff engineer and a list of ideas for tactical initaitives, visit this page.
Team Days
Engineering management schedules Team Days every 2 quarters as a way to celebrate our wins as a team, with our counterparts, and look forward to the future. These are optional social events led primarily by the attendees in terms of activities and discussions, but scheduled and organized by engineering managers.
Sync meetings will occur in all time zones, and activities may happen throughout the day in a dedicated team day Slack channel. While optional, all team members should have an opportunity to participate for as long as they want, so capacity for all engineers should be accounted for during milestone planning.
Metrics
| Dashboards & Handbook | Async Updates | Sync Updates |
|---|---|---|
| Error Budget (Handbook) | Monthly reporting issue | Saas availability when red |
| Past due Infradev (Handbook) | Monthly reporting issue (Foundations, Data Science) | Saas availability when SLA is past or backlog is unmanageable |
| Bug/Vulnerabilities (SLAs) | Monthly reporting issue (Foundations, Data Science) | Saas availability when SLA is breached or backlog is unmanageable |
| Recent Incidents | Reliability Reports | Saas availability when applicable |
| MR Rate (Handbook) | - | Engineering stage review |
| Mean Time to Merge | - | Engineering stage review |
| Review Time to Merge (Handbook) | - | Engineering stage review |
| Review Rate | - | Engineering stage review |
| Technical Roadmap progress (Handbook) (Data Science, Foundations) | - | Engineering stage review |
| Say/Do (Handbook) | Deliverables | Engineering stage review |
Development
Say/Do Ratio
When we commit to a new milestone, engineering managers apply the ~Deliverable label to all issues they are able to commit to based on capacity. This serves as a promise from engineering to product as well as a signal for the most important issues in the milestone for the team to pick up. At the end of the milestone, successfully completed issues with the ~Deliverable label are assessed against the total number of ~Deliverable issues in order to create the Say/Do Ratio (7 issues completed of 10 committed to is a 70% Say/Do result.)
The Say/Do Ratio is a metric used to:
- Provide predictability and stability when milestone planning
- Raise flags about the types of issues that are frequently disruptive
- Measure how well we are meeting our commitments to our customers
Meetings
Although we have a bias for asynchronous communication, synchronous meetings are necessary and should adhere to our communication guidelines. Some regular meetings that take place are:
| Meeting (Internal Only, limited access) | Frequency | DRI | Topics |
|---|---|---|---|
| SaaS Availability | Weekly | Infrastructure | Incident corrective actions, Feature Change Lock status, Past Due Infradev, Error Budget, or Security updates |
| Engineering Stage Review Handbook | Monthly | Michelle Gill | Productivity metrics, Technical Roadmap updates |
| Foundations and Data Science EM Meeting | Biweekly | Michelle Gill | Standards discussions, questions, help needed, FYIs, process improvements, collaborating on engineering stage direction, team event planning, leadership communication dissemination, wider initiative brainstorming, OKRs |
| Core Development Staff Meeting | Biweekly | Tim Zallmann | Recurring staff meeting for updates across engineering management within Tim’s organization |
| AI - Eng Leadership Sync | Weekly | Tim Zallmann | Engineering coordination on shared AI objectives (architecture, inference, prompt migration) across groups and sections |
Stage Reviews
The Engineering Stage Review is a monthly Engineering Manager meeting designed to evaluate team productivity, technical progress, and key metrics across the stage groups. In this meeting, we assess team productivity metrics and trends at a stage-wide level, review technical roadmap progress and align initiatives across teams, identify areas for improvement and discuss action plans, and connect related technical work across different teams.
Format:
- 50-minute monthly sessions, split into two parts:
- Data Science Session (covering AI Framework, Custom Models, MLOps, Duo Chat)
- AI Powered & Foundations Session (covering Duo Workflow, Global Search, Design System, Personal Productivity, Import & Integrate)
- Each team has approximately 10 minutes for review
- Metrics and updates are submitted 24 hours before the meeting for async review
Skip Levels
Skip Levels with Michelle occur in 3 ways:
- Each month, skip level invites will be sent to approximately 10 team members depending on availability across teams. This equates to 1 skip level approximately every 6 months, therefore invites are sent based on the last meeting date.
- Monthly placeholder events are pre-scheduled in all regions also at that time and can be “claimed” on this calendar (internal only) by any engineer or engineering manager. If you don’t see any - they are taken, follow step 3.
- Michelle is never too busy for a skip level. Simply schedule one. 🙂
Links and resources
- Engineering links
- Manage and Data Science engineering discussion are located at
gitlab-org/manage/engineering-managers - Social channel #pikachus-social
- Engineering Manager Slack #s_manage_data_science_ems
- Manage and Data Science engineering discussion are located at
ModelOps
The ModelOps stage aims to empower GitLab customers to build and integrate data science workloads within GitLab.
Last modified February 1, 2025: Add stage review meeting documentation (
05ab939a)
