Senior ML Systems Engineer (C++ & AI Audit)

Full Description

The Challenge

We are looking for a high-caliber Part-time Senior AI Engineer to undertake a deep-dive technical audit of a complex, performance-critical C++ Machine Learning codebase.

This is a "detective" role for an engineer who excels at reverse-engineering undocumented logic and translating low-level C++ implementations into clear architectural blueprints.

Key Responsibilities

• Reverse Engineering: Analyze a legacy C++ ML codebase with minimal documentation to reconstruct its logic.

• Architecture Mapping: Identify and document model architectures, training/inference workflows, and feature engineering pipelines.

• Cross-Language Translation: Map C++ ML concepts (manual tensors, buffers, custom loss functions) to standard Python-based frameworks (PyTorch/TensorFlow).

• System Decomposition: Identify dependencies, data flow patterns, and performance constraints.

• Technical Documentation: Produce a high-quality "Source of Truth" document covering the build environment, model internals, and system limitations.

Required Skills & Experience

• C++ Mastery: Extensive experience in C++ software development, specifically for scientific or performance-critical systems.

• ML Core Theory: Deep understanding of how ML works "under the hood" (gradient descent, backpropagation, and tensor operations implemented from scratch).

• Python ML Ecosystem: Proficiency in Python (PyTorch/TensorFlow) to provide context for C++ translations.

• Code Auditing: Proven ability to navigate, debug, and explain large, unfamiliar codebases.

• Technical Writing: Ability to translate complex code into professional, structured documentation.

Nice to Have

• Experience with C++ ML libraries (e.g., Eigen, ArrayFire, LibTorch).

• Background in MLOps or legacy system migration.

• Experience in research-oriented or experimental codebases.

Comp:

$1000 to $1500 based on skills and experience

Why This Role?

This is a high-impact, independent project ideal for a senior expert who enjoys solving complex puzzles. You will have a direct hand in making a "black box" system understandable and maintainable for the future.