ABAshish Bisht
Resume
ESC
Back to work

Case study · 2025 — Now

Gecko

Chiller‑plant performance service

Gecko serves chiller‑plant performance data to the rest of the suite: live values, part‑load hours, thresholds and bulk exports. I've been one of its main contributors, focused on data correctness, new endpoints and keeping it on current Spring Boot.

My role
Major contributor
Stack
  • Java 17
  • Spring Boot 4
  • JPA
  • PostgreSQL
  • JUnit
  • Mockito

Challenges

  1. 01

    Correct numbers from messy inputs

    Problem
    Equipment data arrives in inconsistent formats. Some values came through as strings or nulls, which led to wrong readings, like severe values showing as zero.
    Approach
    I made parsing defensive and explicit, and added tests for invalid and edge‑case inputs so regressions get caught in CI.
    Outcome
    Live values are reliable, and bad input fails loudly in tests instead of quietly in production.
  2. 02

    Keeping the framework current

    Problem
    Older Spring and Jackson versions were flagged with security issues.
    Approach
    I upgraded to Spring Boot 4 and Jackson 3, fixed the breaking changes, and added SBOM scanning to the pipeline.
    Outcome
    The service runs on supported versions with its dependency risk tracked.

What I took away