Internship Machine Learning Engineer resume example

• Under supervision, versioned training data and model artifacts together, compared offline evaluation with serving latency, and inspected misclassified examples before approving the deployment candidate.
• The team distilled a ranking model and batched feature retrieval for online inference; I supported research preparation, validation, and documentation under mentor review. The team removed the longest wait from the end-to-end workflow; confirmed with my mentor that my contribution was limited to the assigned support and validation work.
• The team added freshness, null-rate, and distribution checks to 62 production features; I supported research preparation, validation, and documentation under mentor review. The team reduced repeat incidents from the same failure mode; confirmed with my mentor that my contribution was limited to the assigned support and validation work.
• Under supervision, owned a PyTorch evaluation slice using frozen data and random seeds, resolving a misleading model gain caused by duplicate examples across train and test sets.
• Under supervision, built a Python deployment check with Kubernetes resource limits, reducing inference failures from 14 to 3 per 1,000 replayed requests at the same load.

• In a mentor-reviewed simulation, built a Python machine learning pipeline using SQL-derived viewing features, keeping training and evaluation windows separate; compared the model against a simple baseline before recommending deployment.
• As a second supervised exercise, added monitoring for prediction freshness, feature drift, and serving errors; tested a fallback ranking path so delayed features would not leave the recommendation surface empty.
• Under supervision, the team built an offline-to-online evaluation suite with slice-level metrics for 14 cohorts; I supported research preparation, validation, and documentation under mentor review
• Generated synthetic source records with duplicates, late arrivals and corrected values; wrote assertions for row counts and key uniqueness, and recorded the expected effect of each case on the reported metric.
• Compared the analytical output with a manually calculated reference table, traced differences to a transformation step, and retained a data dictionary and rerun instructions alongside the corrected query.
• Owned the synthetic-data validation using a manually calculated reference, resolved duplicate-key inflation and completed a notebook that reproduces the corrected totals.

Relevant coursework: Algorithms, operating systems, databases, computer networks

• Published an independent methods note using reproducible queries, explaining the data grain, excluded records and sensitivity of the result to a changed denominator.