Data AI Engineer

About Mindtech

Mindtech is your gateway to exciting, impactful technology projects. We specialize in staff augmentation and end-to-end software outsourcing, connecting top Latin American talent with global opportunities. Our agile approach ensures exceptional service and innovative solutions for our clients.

Were hiring a Data Science Engineer who can bridge:

• Structured datasets with qualitative, unstructured human responses
• Modern LLM tooling (e.g. OpenAI APIs)
• Algorithmic thinking, evaluation, and production delivery

This role sits between data science, applied AI, and engineering.

What Youll Work On

• Analyzing and transforming structured datasets with qualitative human responses
• Designing algorithms that combine historical data + LLM outputs
• Building simulation systems across demographics, intent, and experience
• Working with prompt engineering, tool usage, and multi-step LLM workflows
• Defining evaluation metrics and benchmarking outputs against real data
• Shipping production pipelines in collaboration with product and engineering

Tools & Technologies

Python

SQL / PostgreSQL

LLM APIs

Async processing pipelines

Experimentation and evaluation tooling

What Were Looking For

• 4+ years experience in data science, applied ML, or AI engineering
• Experience working with qualitative, unstructured human-generated data
• Hands-on experience with LLM APIs in applied systems
• Ability to design defensible, benchmarkable algorithms
• Comfort deploying work into production environments
• Strong communication and collaboration skills

Youll Thrive Here If You've

Built applied AI systems used by real users

Combined historical data with model-generated outputs

Worked with evolving specs and ambiguous problem spaces

Enjoy turning qualitative data into reliable, explainable systems

What This Role Is (and Is Not)

This role is:

Applied, analytical, and production-focused

About rigor, evaluation, and real outcomes

This role is not:

Pure prompt engineering

Academic research detached from product

A black-box modeling role without accountability