Maciej Kalisiak
Professional Summary
Generalist SWE with 18 years of Google experience, spanning:
- Solid record as Individual Contributor (L4–L6) and Technical Lead Manager (TLM)
- 3 Google Product Areas: Ads, GKE/Kubernetes, and Gmail
- IC focus: high-reliability, high-scale distributed systems, Observability, Debuggability
- TLM focus: bootstrapping high-performing teams, operational excellence
- AI focus: leveraging LLMs for explaining inherited code, code gen, automating toil
Professional Experience
Gmail Backend Infrastructure
Nov 2022 – Current
- Managed team of 7 in 6mo+ effort modernizing our prod deployment layout.
- Drove Incident Response in a handful of critical (SVP+ visibility) outages; in particular:
- Led team to:
- deploy Google's load-shedding technology across all our key binaries (6)
- simplify approach by turning down & eliminating outdated, complex stack feature
- (IC) Designed & rolled out a quick'n'dirty loadtest framework to reproduce the detected spam attack vector, confirming robustness of our mitigations.
- (IC) Designed & implemented biasing mechanism into our store-and-forward architecture
- Led team to:
- (IC, sideproject) Modernized Gmail's primary (but very aged) debugging tool: rapid implementation in Colab notebook, added new features, Gemini-integration.
- Drove a variety of latency improvements, as well as better measurement; collaborated with our SREs on implementation & debugging of latency SLOs for our service.
- Executed broader team toil reduction initiatives, including tracking & analysis of our Release and OnDuty rotations time spend, retiring low value testing infra, merging rotations.
- Took over & rolled out complex but launch-critical application licensing work; executed across 4+ layers of teams, escalations, debugging of the underlying brittle infrastructure.
GKE / Kubernetes
Jan 2019 – Oct 2022
- Team Growth & Management:
- Instrumental in growth and leadership of the first GCP team in the Waterloo office. GCP now has several hundred engineers in Waterloo.
- Partnered w/Senior Leadership to rapidly grow GKE team in Waterloo.
- Onboarded and ramped-up 2 new managers, off-loaded reports (reached ~17 reports for 5+ months).
- Critical leader who navigated unique challenges: COVID, performance issues, culture friction between employees and externally hired leadership.
- Rapidly built & led a strong team for Anthos Observability in response to sudden mandate; high team promotion rate (~80%).
- Observability Team Leadership:
- Team took on mandate to build and enhance observability through all layers of GKE; built event collection & repository, Datadog-inspired visualization.
- Took over Anthos Observability (monitoring, metrics, and logs) from the disbanded Stackdriver team; reworked as a common solution across OnPrem, OnBareMetal, and OnAWS.
- Grew to include lower-level Kubernetes observability after absorbing scope from sister team.
- Visualization of Cluster History in Pantheon.
Google Ads / Display Ads
Mar 2008 – Dec 2018
- Next-gen Serving System:
- Designed and rebuilt the overall control flow structure (FSM) of central "mixer" service.
- Re-implemented the overall ad targeting flow logic, to parallelize calls to ad exchanges and real time bidders.
- Led design & reimplementation of a more light-weight "targeting signal" computation, to better fit real-time bidder deadlines.
- Owned the entire next-gen launch: experimentation, validation, and production rollout.
- Led team (design, implementation) to devise a turn-key solution for quickly building & introducing further ad provider services (used for AdSense Direct Sales & X-Ads).
- Data Architecture & APIs:
- Co-designed and implemented the API, architecture, plumbing for data about the Publisher.
- Co-designed the unified API for representing which ads won (originally highly heterogeneous, esp. between auction & reservation systems).
- Designed & induced a more structured form for our central ad request/response API (initially mostly flat "kitchen sink").
- Ads Experiments Infra (production-live A/B experimentation system handling trillion events/day): Modernized through flag API update, introduced file-valued flags, better dashboards.
- Skunkworks team: pursued a number of experimental but critical issues:
- Developed emergency release procedure & tooling to bring down cherry-pick binary push process from ~8h to ~1h.
- Tool for correlating prod anomalies with system changes (e.g., binary release, experiments).
- Compliance:
- Designed & implemented frequency capping for mobile ads.
- Ensured COPPA compliance of our team's primary servers.
- Implemented Gambling advertiser blacklisting.
Education
Computer Science | Ph.D. – University of Toronto, 1999–2007
Thesis: Toward More Efficient Motion Planning with Differential Constraints
Publications: list
Computer Engineering | M.A.Sc. – University of Toronto, 1997–1999
Electrical Engineering | B.A.Sc. – University of Waterloo, 1992–1997
Skills
- Languages: Python, C++, Go, SQL; familiar with: Java, Javascript, LISP, Haskell, Lua
- Methodologies: A/B Experimentation, A/B prod diff validation, CI/CD pipelines, Agile, Microservices architecture
- Data analysis: Colab (Jupyter Notebook), R
- Tooling: Google Cloud, Kubernetes, Docker, GitHub, Git, Fig, Linux
- AI: Claude Code, gemini-cli, NotebookLLM
References available upon request.