Experience: 2–3 years (industry or strong internships)
Location: Chennai, India preferred | Remote possible
Role Overview
Mirabilis Design is looking for an AI Engineer to help define and build the AI roadmap for our product and for our customers.
This role is focused on applied AI for engineering productivity, not academic experimentation.
You will work closely with VisualSim modeling engineers to design, train, validate, and deploy AI capabilities that automate and enhance complex engineering workflows.
Key Responsibilities
- Define a roadmap of AI capabilities for internal use and customer-facing features
- Run experiments using existing VisualSim models and data
- Compare AI-generated outputs against expected and validated results
- Identify and define training data requirements
- Work with engineers to generate, clean, and label datasets
- Build and fine-tune LLM-based solutions for engineering use cases
- Design and execute large-scale test scenarios to validate correctness and reliability
Example AI Use Cases
- Generate VisualSim XML models from specification documents
- Generate workloads, configurations, and expected results
- Auto-generate test cases and validation scenarios
- Assist with documentation and model explanations
- Create AI tools that reduce manual engineering effort
Required Qualifications
- Bachelor’s or Master’s degree in:
- Computer Science
- Mathematics
- Statistics
- or a closely related field
- 2–3 years of relevant experience as:
- AI/ML Engineer, OR
- Research/Applied AI Intern with substantial project work
- Strong foundations in:
- Machine Learning
- Statistics and probability
- Data analysis
- Hands-on experience with:
- Python
- ML / AI frameworks (PyTorch, TensorFlow, HuggingFace, etc.)
- LLMs, prompt engineering, or model fine-tuning
Nice to Have
- Experience working with structured or semi-structured data (XML, JSON, logs)
- Exposure to engineering, simulation, or EDA tools
- Experience validating AI outputs against deterministic or expected results
- Ability to work independently and design experiments end-to-end
