AI/Machine Learning Engineering Intern

• Gecko Robotics is seeking an AI/ML Engineering Intern to join our Engineering team in New York for Summer 2027
• Gecko collects some of the richest inspection data in the world: ultrasonic and other non-destructive testing (NDT) measurements from wall-climbing robots, paired with the documents, drawings, and operational records that describe the assets they inspect
• You will work on a defined project that turns that data into decisions, building models and the systems around them that run reliably on real customer data
• This is an engineering internship, not an analytics one. We build the systems that analyze the data, not just the analysis
• Own a scoped machine learning project from problem framing through evaluation and integration, under the guidance of experienced engineers
• Build data pipelines in Python that clean, label, and structure NDT sensor data, documents, or operational records for model training and inference
• Train, evaluate, and iterate on models, and design evaluations that reflect how the output will be used in the field
• Integrate a model or LLM-based workflow into a service or tool that other engineers and Forward Deployment Engineers can use
• Investigate failure modes on messy, real-world data, document what you find, and recommend what to fix next
• Use version control, code review, testing, and experiment tracking to produce work others can reproduce and build on
• Present your project, results, and recommended next steps at the end of the internship
• Technologies We Use:
• Python, PyTorch or similar ML frameworks, NumPy and pandas, SQL, LLM APIs and agent frameworks, Git, Linux, and cloud infrastructure (GCP)
• Depending on the project, you may also work with signal-processing tools for ultrasonic data, document-processing and information-extraction pipelines, or orchestration tools such as Airflow
• The specific tools will depend on the assigned project

Preferred experience

• You are an engineer who happens to work in machine learning
• You care whether a model works on real data, not just on a benchmark, and you are comfortable owning the code around the model: the pipeline, the evaluation, and the integration
• You approach new problems methodically, communicate clearly about results and uncertainty, and are excited to be the first person to look at a problem
• Strong programming experience in Python, including writing well-structured, tested code beyond notebooks
• Ability to work onsite in New York, Monday through Friday, for the full Summer 2027 internship
• Currently pursuing a bachelor’s, master’s, or doctoral degree in computer science, electrical engineering, applied mathematics, physics, or a related technical discipline, with an expected graduation date between December 2027 and May 2028
• Working familiarity with Git or similar version control systems
• Working understanding of linear algebra, probability, and statistics
• Ability to independently close knowledge gaps, stay organized, and learn from feedback
• Coursework or project experience in machine learning, including training and evaluating models on real datasets
• Strong written and verbal communication skills, including the ability to explain results, quantify uncertainty, and collaborate on ambiguous problems
• Graduate-level coursework or research in machine learning, signal processing, or computer vision
• Experience with LLM APIs, retrieval, information extraction, or document understanding
• Experience taking a model beyond a notebook: building a pipeline, an API, or a tool around it
• Experience with time-series or sensor data, especially ultrasonic, acoustic, or other physical measurements
• Interest in industrial, defense, or infrastructure applications where reliability matters
• Exposure to cloud infrastructure, preferably GCP, and to containers such as Docker