W/ William Lu
All cases

Case study

ASML Automation and Leadership Case Study

Read source on GitHub ↗
On this page

The delivery problem

ASML engineers in different time zones shared complex HMI eP5 test environments. A test run required Simulator startup, function-specific Recipe loading, manual operation, and cleanup. Long-running cases kept an engineer at the environment. An incomplete reset could also prevent the next engineer from starting.

I led six Zealogics Automation Engineers in the ASML-side STA team. ASML supplied the base TestComplete automation architecture. I improved its configurability and cross-product use, designed the setup-to-cleanup lifecycle, planned reusable modules, and coordinated delivery across iCDU, MM Flow, and ADEL. I implemented some test cases; the six engineers completed most coding.

Architecture and test strategy

I included the environment lifecycle in the automation architecture.

  • The workflow covered Simulator startup, initialization, retrieval of current test files and test cases, environment setup, Recipe loading, execution, status reporting, cleanup, and reset.
  • Engineers could trigger the normal path from a .bat file or Bamboo, then review the TestComplete or Bamboo result.
  • The normal path ran unattended from setup through reset. A Bamboo connection failure still required an engineer to check the environment and connection state.
  • I reviewed test cases before assignment. I grouped repeated behavior into reusable modules, then divided implementation work among the six engineers.
  • I assessed whether each test case supported full automation. When it did not, I proposed partial automation to remove repeatable manual work while keeping the required human step visible.
  • Bitbucket held source code. Bamboo and TestComplete handled CI test orchestration and result review. The team also used Python, VS Code, Jira, and Confluence.

Delivery and team leadership

I planned the technical work and tracked progress, risks, and requests for help across teams.

  • Daily syncs surfaced environment blockers, dependencies, and work that needed help from another ASML function owner.
  • I represented STA in management and weekly status meetings. Jira filters and JQL supported ticket and test case status reviews, but the original queries and counts are no longer available.
  • I used one-on-ones for blockers, support, and direct performance feedback. Annual formal evaluations went to my direct manager; the later approval chain is unknown.
  • I reviewed the delivery scope, assigned tasks and test cases, discussed implementation details with the engineers, and tracked risk, schedule, tickets, and handoff.
  • I created Jira and Confluence material for test cases, Test System Design, work records, operation guidance, and newcomer onboarding. I cannot confirm whether the onboarding material remained in use after my engagement.

Training and feedback

I provided training, live demonstrations, operating instructions, and technical documentation to client engineers across functions. Approximately 20–30 people attended.

I can confirm that approximately 2–3 engineers ran the automation independently after training and submitted change requests. I do not know how many of the other attendees used it. The team discussed informal feedback in daily syncs or Communicator and tracked formal requests in Jira. I completed what the schedule allowed; ASML inherited the open requests after the engagement ended.

Handoff during the budget change

ASML reduced its software-outsourcing budget, and STA lost funding. I documented the six engineers' capabilities, arranged work handoffs, and worked with my direct manager on reassignment. All six moved to other ASML teams or projects. Engineering leadership and HR made the final placement decisions.

Recorded outcome

  • Engineers could start the normal workflow through a batch file or Bamboo and review one result after unattended setup, execution, cleanup, and reset.
  • The workflow reduced repeated manual setup, operation, and cleanup work. It also gave users a consistent reset path for shared environments.
  • Reusable modules and pre-assignment review reduced duplicate implementation across the three HMI eP5 modules.
  • Independent use and Jira feedback provided limited adoption evidence. Session attendance alone was not counted as adoption.
  • ASML received the technical material, open Jira requests, and team handoffs when the engagement ended.

Quality and leadership decisions

  1. Include the environment lifecycle in the architecture. Automation that leaves setup or reset to memory can still block the next user.
  2. Choose full or partial automation case by case. A visible manual boundary is safer than claiming unattended execution for a step that still needs judgment.
  3. Separate architecture ownership from coding volume. I designed and coordinated the improvements; six engineers implemented most code.
  4. Separate attendance from use. Training attendance shows who took part. Independent runs and modification requests show stronger adoption evidence.
  5. Keep exception and ownership boundaries explicit. Connection failures required manual investigation, and ASML leadership and HR owned final staffing decisions.

Reusable controls

  • Model setup, data retrieval, execution, reporting, cleanup, and reset as one stateful workflow.
  • Add health checks for environment availability, connection state, required files, services, and final reset state.
  • Review test cases before assignment and extract reusable modules before parallel implementation begins.
  • Record whether a test case is fully automated, partially automated, or manual, including the reason and human handoff point.
  • Track delivery status, environment blockers, feedback, and open requests in one reviewable ticket flow.
  • Keep training attendance, independent use, feedback, and long-term adoption as separate measures.
  • Prepare technical documentation, capability records, and open-work ownership before a team transition.

Evidence limits

  • This case is based on my direct work and recollection. Original project records are no longer accessible.
  • Setup-time, cost, and coverage improvements are described without percentages because the source measurements are unavailable.
  • The case study does not claim that I created ASML's base architecture, implemented the full system alone, automated every test case, prevented downtime, or delivered zero-failure automation.
  • Bitbucket use is limited to source control. No Git workflow, branch strategy, pull-request ownership, Bamboo integration detail, or separate reporting tool is claimed.
  • Individual performance details, Jira queries, ticket counts, and post-departure use of onboarding material remain private or unknown.

Period: May 2023–Apr 2024
Role: Project Leader / Senior Automation Test Engineer, Zealogics／ASML engagement
Scope: HMI eP5 modules iCDU, MM Flow, and ADEL; fully remote distributed team
Public navigation: Portfolio | Public QA resume

Source: Based on my direct work and recollection. Original employer records are no longer available for independent verification. Confidential details are omitted.