Senior Software Engineer, Data Science (AI Accuracy)

About DroneDeploy

DroneDeploy provides field teams with robotics and AI to help critical industries work faster and with greater confidence. Its reality-capture platform combines robotic automation, AI agents, drones, 360 and fixed cameras, ground robots, and proprietary AI so customers can improve safety, simplify field operations, and make better decisions from the field to the boardroom.

The company operates in a mostly remote-first environment built around innovation, impact, and collaboration. It is known for a supportive team culture, meaningful technical work, flexible schedules, family-friendly benefits, and opportunities to grow through internal promotion.

Role overview

This six-month, full-time contract role is centered on improving the trustworthiness of DroneDeploy’s AI systems. As a Senior Software Engineer, Data Science on the AI Accuracy team, you will own how accuracy is measured, analyzed, and improved for LLM-based features across Progress AI and related products.

You will create and maintain evaluation datasets and pipelines, build internal tooling for audits and error analysis, and help refine prompts and underlying models from experimentation through production rollout and ongoing monitoring. The role connects ontology and schema design with real-world accuracy outcomes, ensuring that changes to taxonomies and data flow cleanly through models, evaluations, and tools.

You will work with a high level of ownership, prioritize work based on impact, and collaborate closely with engineering, product, and Inspection/Progress AI stakeholders. The position reports to the Senior Manager of Software Engineering for AI Insights/Inspection AI.

Work environment

• Based in Auckland, New Zealand, with a strong preference for someone who can work regularly from the Auckland office alongside the Inspection and Progress AI teams.
• Hybrid working is supported, with at least two in-office days each week.
• Regular overlap with Auckland business hours, roughly 9:00 a.m. to 5:00 p.m., is expected to support collaboration with local and global colleagues and occasional on-call duties.
• Occasional domestic and international travel for team onsite sessions may be required.
• AI-assisted tools are part of the engineering workflow for coding, data exploration, prompt iteration, and log/debug analysis, while quality, safety, and maintainability remain the engineer’s responsibility.

Responsibilities

• Develop and support dependable dataset and evaluation systems, including controls that safeguard ground-truth quality.
• Investigate and resolve gaps between offline evaluation results and real production accuracy.
• Improve and expand labeling frameworks and knowledge-capture processes.
• Manage prompt optimization from shortlisting candidates through rollout and post-launch validation.
• Automate evaluation pipelines and build scalable tooling for use across multiple AI products.
• Create reusable internal tools for error analysis, audits, and experimentation, while focusing effort on the highest-impact work.

Requirements

• Strong hands-on experience with Python for data manipulation, analysis, and tooling, including libraries such as pandas and NumPy.
• Solid SQL and database skills, including querying, data modeling, and handling large datasets.
• Background in data science with knowledge of LLM accuracy, especially visual reasoning in these models.
• Experience creating and maintaining evaluation datasets, debugging evals with tools such as Braintrust or Langfuse, and performing error analysis.
• Ability to design ground-truth and labeling systems, including measuring label agreement and data quality.
• Understanding of prompt engineering and the full prompt lifecycle from candidate selection to deployment and validation after launch.
• Capability to reason about ontology and schema design and how those choices affect models, evaluations, and tooling.
• Domain familiarity with construction or industrial inspection is an advantage.
• A collaborative, iterative, team-oriented working style.
• Comfort using AI-assisted tools for coding, analysis, and debugging, with strong engineering judgment to avoid hallucinations and maintain code quality.
• Drone certification is not needed for this position.

Additional information

DroneDeploy is an equal opportunity employer. All employees are expected to protect company and customer data by following information security policies and procedures. The recruitment process may use AI tools to help review applications, analyze resumes, and check responses for potential inconsistencies or verification signals, but final hiring decisions are made by humans.

Candidates from all backgrounds are encouraged to apply, even if they do not meet every listed requirement. Applicants are invited to share the strengths and experience they would bring to the AI accuracy work.

Location and schedule notes

• Location: Auckland, New Zealand.
• Work model: hybrid, with an in-office expectation of at least two days per week.
• Business-hour overlap: regular availability during standard Auckland hours is required.
• Travel: occasional domestic and international travel may be part of the role.

Privacy and compliance

For details on data handling during recruitment, candidates should review the company’s recruitment privacy notice. If you want more information about how application data is processed, the company can be contacted directly.