Sensor Calibration Assistant with Google ADK
Field technicians calibrating complex chemical sensors like Emerson's Rosemount 350A need step-by-step interactive verification tools. In this challenge, developers will use Google ADK to build a light, voice- or text-driven calibration assistant that guides operators through buffer immersion and zero-point offset steps.
What you are building
The core problem, expected build, and operating context for this challenge.
Construct a step-by-step sensor calibration guide using Google ADK to assist operators with Rosemount 350A pH sensors.
How work is evaluated
Evaluates calculation of electrode slope percentage based on test mV inputs.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
Slope Accuracy Test
Slope > 95% must return pass_status = True
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Math Accuracy Score
Precision of slope calculation against theoretical Nernst equation value • target: 1 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Configure a Google ADK agent for step-by-step procedural workflows
Parse operator millivolt (mV) readings during pH 4, 7, and 10 buffer tests
Calculate sensor slope percentage and flag degraded glass electrodes
Output calibration certificate documentation
Reference links and supporting material
Synthetic pH sensor calibration test records and buffer voltage measurements.
Challenge setup
Evaluates calculation of electrode slope percentage based on test mV inputs.
Challenge input
JSON with mv_ph4, mv_ph7
Google ADK
Assigned framework
Evaluated output
JSON with calculated_slope_pct, pass_status (boolean)
- Slope > 95% must return pass_status = True
- Precision of slope calculation against theoretical Nernst equation value • target: 1 • range: 0-1
- Benchmark: Industrial Math & Calculation Benchmark
- Math Accuracy Score target: 1
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Google · Required
- Google ADK · Optional
- Google Jules · Optional
- Google Jules · Optional
Evaluation contract
- Slope Accuracy Test · Weight 1
- Math Accuracy Score · Weight 1
Metrics
- Math Accuracy Score · Target 1
Benchmarks
- Industrial Math & Calculation Benchmark
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON with mv_ph4, mv_ph7
Output: JSON with calculated_slope_pct, pass_status (boolean)
Execution harness: Python
Evaluation tools
- Google ADK · Agent development kit
Recipe freeze marker
No recipe freeze time is recorded.