Sania Qamar's Optum ML Assessment Experience on CodeSignal

Source: medium.com

TL;DR

• Sania Qamar shares her experience taking Optum's Machine Learning Core Assessment on CodeSignal for a Senior AI/ML Engineer role.
• The test lasted 1 hour 10 minutes with 6 quiz questions (some multiple correct answers) and 3 coding questions.
• It tests ML theory, metrics like recall and false positive rate, numerical reasoning, and coding core algorithms from scratch.[[1]](https://medium.com/@sq.sania/my-experience-with-optums-machine-learning-core-assessment-for-a-senior-ai-ml-engineer-role-33c25d111175)

The story at a glance

Sania Qamar, a senior data scientist, describes her recent attempt at Optum's Machine Learning Core Assessment on the CodeSignal platform as part of interviewing for a Senior AI/ML Engineer position. The assessment evaluates theoretical ML knowledge alongside practical skills like implementing algorithms under time pressure. It is being shared now to offer a practical breakdown for others preparing for similar senior-level tests, though full details beyond the intro and first quiz are paywalled.[[1]](https://medium.com/@sq.sania/my-experience-with-optums-machine-learning-core-assessment-for-a-senior-ai-ml-engineer-role-33c25d111175)

Key points

• Total duration: 1 hour 10 minutes, split into 6 quiz questions and 3 coding questions focused on core ML algorithms.
• Quiz questions often have multiple correct answers, demanding careful option review; designed for theoretical knowledge and numerical reasoning.
• First quiz: Analyze four confusion matrices (some with 0/1 labels swapped) to find the model with Recall > 90% and False Positive Rate < 10%; requires identifying TP/FP/FN/TN and mental calculations.
• Quiz difficulty high due to tricks like label flips, stressing double-checking metric definitions and interpretations.
• Overall aim: Gauge practical ML understanding beyond theory, including coding implementations from scratch.
• Promises coverage of quiz concepts, coding challenges, and key takeaways, but further specifics not visible publicly.[[1]](https://medium.com/@sq.sania/my-experience-with-optums-machine-learning-core-assessment-for-a-senior-ai-ml-engineer-role-33c25d111175)

Details and context

Optum, part of UnitedHealth Group, uses this assessment early in hiring for senior AI/ML roles to filter candidates quickly. CodeSignal is a common platform for timed coding and quiz tests in tech interviews, emphasizing no external libraries for core ML tasks.

The visible first quiz highlights real-world evaluation challenges: label inconsistencies mimic production data issues, testing deep metric familiarity over rote recall.

No outcomes, other quizzes, coding details, or preparation tips are accessible; secondary sources like Reddit mention Optum online assessments but lack matching specifics on this test.[[2]](https://www.reddit.com/r/learnmachinelearning/comments/1qoljly/anyone_interviewed_for_ml_engineer_at_uhgoptum)

Key quotes

None available from visible content.

Why it matters

Senior ML interviews increasingly blend theory, metrics, and timed coding to ensure production-ready skills amid rising AI hiring demand. Job seekers targeting Optum or similar firms gain a preview of format and trick questions, helping prioritize prep on fundamentals like confusion matrices. Watch for author updates or shared experiences on forums, as full paywalled details could reveal more on coding problems.[[1]](https://medium.com/@sq.sania/my-experience-with-optums-machine-learning-core-assessment-for-a-senior-ai-ml-engineer-role-33c25d111175)