Deep Learning Researcher

Graph Neural Network & Reinforcement Learning Engineer

Location:

Job Type:

Experience Required:

About the Role

Key Responsibilities

• Design and train Graph Convolutional Neural Network architectures for spatial reasoning tasks
• Develop and optimize Reinforcement Learning agents for constraint satisfaction problems
• Build robust data preprocessing pipelines for graph-structured engineering data
• Implement multi-objective optimization frameworks with competing constraints
• Validate model performance on real-world design problems and production datasets
• Collaborate with domain experts to translate engineering requirements into ML objectives
• Deploy trained models in production environments with performance monitoring

Must-Have Skills

• 5+ years training GCNN models

   using PyTorch Geometric, DGL, or similar frameworks

• Proven experience with RL algorithms

  : Policy Gradient, Actor-Critic, Q-learning implementations

• Graph data engineering expertise

  : adjacency matrices, node/edge features, graph sampling techniques

• Data curation skills

  : cleaning, augmentation, and preprocessing of structured engineering data
• Production ML experience: model versioning, A/B testing, performance monitoring

• Optimization experience

  : multi-objective problems, constraint handling, convergence analysis
• Proficiency in 

  Python, PyTorch/TensorFlow, and distributed training

   setups

Graph-Specific Training Experience

• Experience training on 

  large-scale graphs

   (10K+ nodes) with efficient batching strategies
• Knowledge of

  graph convolution variants:

  GCN, GAT, GraphSAGE

• Graph augmentation techniques

  : node dropout, edge perturbation, subgraph sampling
• Experience with 

  heterogeneous graphs

   and multi-relational data
• Understanding of 

  graph pooling

   and hierarchical graph representations

RL Training Experience

• Environment design

   and custom reward function engineering
• Experience with 

  continuous and discrete action spaces

• Curriculum learning

   and progressive difficulty in training

• Distributed RL training

   and parallel environment execution
• Knowledge of 

  exploration strategies

   and handling sparse rewards

Preferred Qualifications

• M.Tech/PhD in Computer Science, Machine Learning, or related field
• Experience with 

  spatial optimization

   or

  placement problems

• Background in 

  constraint satisfaction

   or 

  combinatorial optimization

• Prior work on

  engineering automation

What We Offer:

• Opportunity to apply cutting-edge ML to solve real engineering problems
• Direct impact on hardware design industry transformation
• Flexible work arrangements and competitive compensation
• Equity participation in high-growth startup