Applied Machine Learning Research Engineer Intern

Quick Facts

• Role: Applied Machine Learning Research Engineer Intern

• Focus: Game Portability & Runtime Control

• Project length: 8–12 week feasibility research (start date and weekly hours agreed before participation)

• Compensation: Unpaid (no promised salary, stipend, equity, or conversion to a paid role)

Description

You will investigate whether learned representations and planning methods can improve game portability and runtime control in real systems settings. You will turn research questions into reproducible experiments, train and evaluate models, and compare results on unseen cases with clear success/failure criteria. A well-supported negative result is explicitly valued.

Responsibilities

• Convert a narrowly defined research question into reproducible experiments with success/failure criteria

• Curate versioned datasets from approved test cases, execution traces, and measurements; prevent data leakage between training and evaluation

• Establish strong baselines before implementing experimental model representations or components

• Train, debug, and evaluate models in Python/PyTorch within an agreed compute budget

• Test generalization on withheld game/hardware pairs and dependency configurations, including failure and recovery cases

• Compare accuracy, constraint violations, latency, memory use, and experiment cost at comparable resource budgets

• Document code, configurations, results, and limitations for reproducibility

• Work on a focused research prototype (not full ownership of product safety architecture, hardware infrastructure, or game-porting pipeline)

Benefits

• Experience building reproducible ML research systems and evaluation protocols for real-world problems

• Deliverables include a reproducible repository, documented datasets/evaluation splits, an experiment log, and a research prototype suitable for engineering review

Requirements

• Hands-on Python and PyTorch experience (training/debugging beyond calling a hosted API)

• Evidence of reproducing a paper, building an experimental ML system, or completing a substantial research project

• Understanding of evaluation design: held-out data, ablations, fair baselines, and sources of misleading results

• Ability to operate within GPU memory/compute constraints and explain tradeoffs

• Clear written reasoning with reproducible code and willingness to report contradictory results

• Useful but not required: transformers, learned discrete representations, vector quantization, compilers/graphics/compatibility systems, systems telemetry

• Complete four short application questions (repo/paper/project link; leak-free evaluation plan; a debugging story; availability and clarification needs for the unpaid internship)