Machine Learning Engineer - Engineering Models

Ubifly Technologies; Chennai District; On-site; INR 1,200,000 - 2,400,000; Full time

Job summary

The ePlane Company, incubated at IIT Madras, is building the world's most compact electric flying taxi and is seeking a seasoned ML engineer to join the data-driven product team in Chennai.

You will design data audits, build pipelines for model training and validation, and deploy AI tools to engineering teams, helping advance surrogate modeling and physics-informed approaches for real-world aviation simulations.

Qualifications

• 3+ years ML engineering with focus on deep learning for scientific or engineering apps.
• Experience training regression/emulation models on physics or simulation data.
• Strong ML stack: PyTorch or TensorFlow, Pandas, NumPy, SciPy.
• Surrogate modeling via Neural Networks or Gaussian Processes for fast proxies.

Responsibilities

• Conduct systematic data audits of existing simulation data, assess schema, volume, cleanliness, gaps.
• Define supplementary data generation requirements for model training.
• Build and maintain data pipelines for model training, validation, and retraining.
• Develop training pipelines, architecture, and prototyping for ML algorithms.
• Work on productising research prototypes.
• Conduct experiments to benchmark techniques and evaluate model behavior.
• Develop evaluation methodology: test sets, accuracy metrics, citation quality scoring, false positives/negatives.
• Deploy AI tools to engineering teams with pilots, baseline measurement, and adoption results.

Skills

ML engineering

Data pipelines

Surrogate modeling

Python programming

Deep learning

Tools

PyTorch

TensorFlow

Pandas

NumPy

SciPy

Gaussian Processes

Job description

Chart the Course for the Future of Flight

The ePlane Company is at the forefront of India's urban air mobility revolution. Incubated at IIT Madras, we are a deep-tech startup dedicated to designing and building the world's most compact electric flying taxi. Our mission is to make door-to-door flying a reality, drastically reducing commute times and decongesting our cities for a cleaner, greener future. We're a passionate team of engineers, designers, and visionaries working on cutting-edge technology, and we're looking for brilliant minds to help us take flight.

Roles and Responsibilities

• Conduct systematic data audits of existing simulation data including schema assessment, volume, cleanliness, and gaps; define supplementary data generation requirements
• Build and maintain data pipelines for model training, validation, and continuous retraining
• Build multi-domain model pipelines that chain individual surrogate models without manual handoff
• Develop training pipelines, architecture, and prototyping for ML algorithms
• Work on productising research prototypes
• Conduct experiments to benchmark new techniques and evaluate model behavior
• Develop systematic evaluation methodology: test sets, accuracy metrics, citation quality scoring, false positive/negative analysis
• Deploy AI tools to engineering teams with structured pilots, baseline measurement, and documented adoption outcomes

Requirements

Required Qualifications

• 3+ years ML engineering with a focus on deep learning for scientific or engineering applications
• Experience training regression/emulation models on physics or simulation data (surrogate modelling or reduced order modelling)
• Strong ML stack: PyTorch or TensorFlow, Pandas, NumPy, SciPy
• Surrogate modeling via Neural Networks or Gaussian Processes for use as fast-running model proxies.
• Proven understanding of fundamental data structures and the ability to apply them to solve complex problems.
• Development experience with retrieval pipeline skills and relational databases

Preferred Qualifications

• Understanding and deployment of Reinforcement Learning based tools
• Understanding of mathematics, particularly linear algebra and probability theory
• Experience with physics-informed neural networks (PiNNs) or hybrid physics-ML models
• Experience with multi-fidelity modelling or chained model pipelines
• Modeling complex multi-physics systems of ODEs and DAEs
• Gradient-based optimization
• Automatic differentiation tools and development