Senior SDET · AI Quality Engineer

Sai Vamsi Kolla

Quality engineering for AI systems, APIs, microservices and UI automation.

I test how AI systems, APIs and web applications behave under realistic conditions. I design automation, model difficult data and failure paths, improve testability and use logs, traces, browser and network evidence to diagnose failures.

View Resume View engineering projects

Vancouver, BC, Canada. Canadian citizen. Eligible to apply for TN status for qualifying U.S. roles.

LinkedIn GitHub Sai Vamsi Kolla
Portrait of Sai Vamsi Kolla

Building confidence in complex systems

  • 7+ years Quality engineering
  • 20 QA engineersCoordinated across distributed teams
  • Building nowTest automation frameworks for AI, APIs and web apps

What I engineer

Most of my work falls into three areas: testing AI behavior, testing APIs beyond the happy path, and building reliable browser automation.

AI systems and behavioral quality

I test whether AI systems understand user intent, use tools correctly, follow expected rules and produce useful results. I also look at what happens when information is missing, ambiguous or incorrect.

  • User intent and expected behavior
  • Tool-use validation
  • Output quality and failure handling
  • Testing non-deterministic systems

API, microservices and realistic data

I test APIs and service boundaries beyond the happy path with realistic data, integration scenarios and deliberately difficult inputs.

  • API contracts and schema validation
  • Integration and fuzz testing
  • Realistic and stateful test data
  • Missing, extreme and invalid inputs

UI automation, testability and observability

I use Playwright to build browser tests that are stable enough to trust in CI. When a test fails, I look at the network, console, traces and application state instead of relying on screenshots alone.

  • Semantic locators and reusable fixtures
  • Network, console and trace debugging
  • Accessibility and responsive checks
  • CI smoke and regression tests

Engineering projects

Building now

These are active projects I use to build and experiment with AI, API and browser testing.

Active build

qa-agent

Evaluation baseline implemented

Problem
Small local language models can sound confident while producing inconsistent or unsupported testing guidance.
Implemented
I built a local RAG-based testing mentor with document ingestion, grounded retrieval, citations, abstention rules and an evaluation harness. The current work is focused on improving retrieval quality and making the evaluations more reliable.
  • Python
  • pytest
  • ChromaDB
  • DeepEval
View qa-agent on GitHub

Active build

fhir-contract-fuzz

Contract testing foundation implemented

Problem
APIs can behave correctly with normal data but fail when requests contain missing, unusual or invalid values. These failures are difficult to discover with happy-path tests alone.
Implemented
I built the first contract and fuzz-testing setup to send both normal and unusual API inputs and validate the responses. The project is now expanding into more workflow and failure scenarios.
  • pytest
  • Schemathesis
  • Pydantic
  • Docker
View FHIR project on GitHub

Active build

medusa-checkout-e2e

CI foundation implemented

Problem
Checkout spans storefront, backend, data and third-party boundaries where timing and state make brittle UI tests easy to create.
Implemented
I set up the Medusa application in Docker and built the first Playwright test structure around it. The project currently has reusable page abstractions, controlled test data, startup checks, smoke tests, Allure reporting and CI. I’m now expanding the cart and checkout scenarios.
  • Playwright
  • JavaScript
  • Docker
  • Allure
View Medusa project on GitHub

Experience and leadership

Quality ownership across AI products, digital commerce, insurance systems and complex enterprise platforms.

Recent work is organized around the systems, risks and responsibilities I owned, not a repeated resume inventory.

  1. 2026 to presentVancouver

    mimik

    Quality Engineer

    Own quality engineering for a multi-agent AI product across AI behavior, APIs, microservices, data flows and end-to-end product risk.

    • Defined the test strategy and quality process for a new AI product area.
    • Built API and service automation using Schemathesis and pytest.
    • Designed tests for expected AI behavior, structured outputs, difficult data and cross-service failures.
    • Improved testability through logging, reporting, tracing and deterministic test data.
  2. 2025 to 2026Vancouver

    HCLTech · Client: Meta

    Test Lead

    Led functional quality work for AI-assisted commerce workflows across 10–15 merchant integrations on iOS, Android, and Shopify.

    • Coordinated the work of 20 offshore QA engineers.
    • Aligned product, engineering, and QA around integration risk and release readiness.
    • Supported quality decisions across multiple merchant and platform boundaries.
  3. 2022 to 2024Edmonton

    Peace Hills General Insurance

    Senior QA Analyst

    At Peace Hills, I tested insurance workflows across Guidewire ClaimCenter, PolicyCenter and BillingCenter, including the APIs and integrations connecting them to other systems.

    • Built the team’s first Selenium and Python regression automation framework.
    • Worked across API, database, integration and functional testing.
    • Tested complex claims and insurance workflows across connected systems.

Earlier experience

QA Analyst · TELUS2021 to 2022

System integration, regression and A/B validation across 12 Agile releases.

Software Engineer, QA · Accenture2016 to 2019

Supported enterprise Guidewire implementations through UI, API, database, integration and release validation.

Engineering approach

Four steps from intent to confidence.

A concise loop for making product risk, system behavior and failure evidence understandable.

  1. 01

    Understand intent and product risk

    Clarify what the system must accomplish and where behavior, data, integration, accessibility or reliability can fail.

  2. 02

    Model realistic data and failure paths

    Create valid, missing, extreme, contradictory, invalid and stateful scenarios.

  3. 03

    Validate contracts and behavior

    Check service boundaries and deterministic results, then evaluate AI behavior against clear expectations.

  4. 04

    Diagnose failures and communicate confidence

    Connect logs, traces, browser evidence and system state to explain failures, coverage and remaining risk.

Credentials

Foundations that support the engineering work.

Advanced automation, AI, cloud foundations, insurance systems and software engineering education.

Certifications

  • ISTQB Advanced Test Automation Engineer2025 · Advanced
  • AWS Certified AI Practitioner2024 · Early adopter
  • Guidewire Certified Associate2024 · Insurance Suite Analyst
  • ISTQB Certified Tester2018 · Foundation level

Education

  • Master of Science, Internetworking TechnologiesUniversity of Alberta · 2021
  • Bachelor of Technology, Information TechnologyGITAM University · 2015

About

Adaptability grounded in ownership.

I moved from India to Canada in 2019 for my master’s degree and progressed from dishwasher to front supervisor while studying.

That experience strengthened the adaptability, ownership, calm coordination and persistence I now bring to quality engineering and technical leadership.

Let’s work together

Need complex systems to be testable and explainable?

I work on AI systems, build reliable test automation, create realistic test data and make failures easier to diagnose.

Email

saivamsikolla@gmail.com Sai Vamsi Kolla
Location
Vancouver, BC
Availability
Canada and US opportunities