Lead AI Engineer

Weekday AI (YC W21); Bengaluru; On-site; INR 3,000,000 - 7,000,000; Full time

Job summary

Weekday AI (YC W21) in Bengaluru is seeking a highly experienced AI Research Engineer to join an advanced team focused on Vision-Language Models, diffusion models, and multimodal AI for real-time, edge-enabled systems.

You will shepherd research from problem definition through prototyping, evaluation, and production hand-off, building scalable pipelines and optimizing models for Jetson-class hardware using CUDA, TensorRT, and ONNX Runtime.

Qualifications

• 8+ years of deep-learning research and development.
• Hands-on experience with diffusion models (DDPM, LDM, ControlNet).
• Experience with multimodal transformers and Vision-Language Models (CLIP, BLIP-2, LLaVA, Flamingo).
• Proficiency in Python and PyTorch; scalable model training and experiment tracking.

Responsibilities

• Research and develop diffusion-based generative models for photorealistic surface simulation, defect synthesis, and domain adaptation.
• Design and train VLM and VLA architectures that integrate textual instructions, CAD plans, visual inputs, and sensor data.
• Develop scalable data-centric AI pipelines using active learning, pseudo-labeling, self-training, weak supervision, and synthetic data.
• Build and optimize deep-learning models for large-scale training and real-time inference.
• Apply techniques such as INT8 quantization, LoRA, knowledge distillation, and model compression for edge deployment.
• Optimize models for Jetson-class hardware, CUDA, TensorRT, and ONNX Runtime within robotics environments.
• Own the complete research lifecycle, including problem definition, literature review, prototyping, experimentation, evaluation, and production hand-off.
• Establish offline and online evaluation frameworks to measure model accuracy, robustness, latency, and scalability.
• Collaborate with perception, robotics, controls, and engineering teams to integrate AI solutions into production systems.
• Prepare internal technical reports and contribute to external research publications and conferences.
• Mentor interns and junior AI/ML engineers and provide technical leadership on research initiatives.

Skills

Diffusion models

Vision-Language Models

Multimodal transformers

Python

PyTorch

CUDA

Edge deployment

Model compression

Research leadership

Education

MS/PhD in CS/EE/Robotics

Tools

TensorRT

ONNX Runtime

DeepSpeed

Ray

Job description

Job Description:

This role is for one of the weekdays clients

Salary range: Rs 3000000 - Rs 7000000 (ie INR 30-70 LPA)

Experience: 8+ yrs

Location: Bengaluru

Job Type: Full-time

We are looking for a highly experienced AI Research Engineerto join an advanced AI research team focused on transforming cutting-edge developments in Vision-Language Models (VLMs), Vision-Language Action Models (VLAs), diffusion models, and multimodal AI into robust, real-time systems for dynamic construction environments.

This role offers an opportunity to work at the intersection of deep learning research, computer vision, generative AI, robotics, and edge deployment. You will be responsible for taking research concepts from problem definition and experimentation through evaluation and production hand-off. The ideal candidate will have strong research depth, hands-on experience with large-scale AI systems, and the ability to translate complex mathematical and theoretical concepts into reliable production solutions.

• Research and develop diffusion-based generative models for photorealistic surface simulation, defect synthesis, and domain adaptation.


• Design and train VLM and VLA architectures that integrate textual instructions, CAD plans, visual inputs, and sensor data.


• Develop scalable auto-annotation and data-centric AI pipelines using active learning, pseudo-labeling, self-training, weak supervision, and synthetic data.


• Build and optimize deep-learning models for large-scale training and real-time inference.


• Apply techniques such as INT8 quantization, LoRA, knowledge distillation, and model compression for edge deployment.


• Optimize models for Jetson-class hardware, CUDA, TensorRT, and ONNX Runtime within robotics environments.


• Own the complete research lifecycle, including problem definition, literature review, prototyping, experimentation, evaluation, and production hand-off.


• Establish offline and online evaluation frameworks to measure model accuracy, robustness, latency, and scalability.


• Collaborate with perception, robotics, controls, and engineering teams to integrate AI solutions into production systems.


• Prepare internal technical reports and contribute to external research publications and conferences.


• Mentor interns and junior AI/ML engineers and provide technical leadership on research initiatives.

• 8+ years of experience in deep-learning research and development, or an advanced degree such as an M.S./Ph.D. in Computer Science, Electrical Engineering, Robotics, or a related discipline.


• Strong hands-on expertise in diffusion models, including DDPM, LDM, and ControlNet.


• Experience with multimodal transformers and Vision-Language Models, such as CLIP, BLIP-2, LLaVA, or Flamingo.


• Proven experience building large-scale, data-centric AI workflows involving active learning, pseudo-labeling, weak supervision, or synthetic data.


• Advanced proficiency in Python and PyTorch or JAX, along with experience in scalable model training and experiment tracking.


• Familiarity with PyTorch Lightning, DeepSpeed, Ray, or comparable distributed-training frameworks.


• Strong understanding of CUDA, C++, TensorRT, and ONNX Runtime for performance optimization and edge AI deployment.


• Solid mathematical foundation in probability, optimization, information theory, and machine learning.


• Ability to translate advanced research concepts into clean, scalable, production-ready implementations.


• Strong problem-solving, research, communication, and technical leadership skills.


• Experience with ROS 2, Nav2, MoveIt 2, Open3D, or robotics perception systems is an added advantage.


• Exposure to synthetic data generation using platforms such as Isaac Sim will be highly valuable.

Requirements: