Edge AI Engineer

Vedya Labs; Hyderabad, Bengaluru; On-site; INR 900,000 - 1,500,000; Full time

Job summary

Vedya Labs in Hyderabad, India is seeking an experienced embedded AI/ML engineer to design and optimize AI models for edge devices. You will work with GPU/DSP/NPU accelerators, tune inference pipelines for memory and power efficiency, and integrate optimized models into various runtimes.

You will leverage CNNs, transformers, and quantization techniques across Linux/RTOS/bare-metal environments and collaborate across teams to ensure high-quality software delivery and robust performance.

Qualifications

• 4–8 years of hands-on embedded AI/ML development in a related engineering field.
• Proficiency in C/C++ and Python with performance-oriented coding practices.
• Experience with heterogeneous compute platforms (CPU+GPU/DSP/NPU) and exposure to DSPs/NPUs such as ARM-NEON, TI C6x/C7x, Tensilica Vision, CEVA, or Hexagon HVX.
• Strong knowledge of processor/SoC architecture, VLIW/SIMD, DMA and memory hierarchy.
• Hands-on experience with CNNs, transformers, and quantization on embedded systems.

Responsibilities

• Design, develop and optimize AI/ML models for edge devices using hardware accelerators.
• Analyze performance, identify bottlenecks, and tune inference pipelines for memory, power and compute efficiency.
• Integrate optimized models into inference runtimes such as TensorRT, ONNX, TFLite, SNPE, OpenVINO or TVM.
• Apply quantization, pruning and other model compression techniques.
• Support integration of AI workloads into embedded software stacks running on Linux, RTOS or bare-metal systems.
• Contribute to SDLC activities: requirements, design, test plans, reviews.

Skills

C/C++

Python

Edge AI

Model optimization

CNN

Transformers

Quantization

TensorRT

ONNX

TFLite

TVM

OpenVINO

DSP/NPUs

GDB/JTAG

Git/SVN/JIRA

Education

Bachelor's or Master's degree in Engineering

Tools

TensorRT

SNPE

TFLite

TVM

OpenVINO

Job description

Role & Responsibilities:

• Design, Develop and Optimize AI/ML models (e.g., CNNs, transformers) for edge devices using hardware accelerators like GPU, DSP, and NPU.
• Analyze performance & identify bottlenecks and tune inference pipelines for memory, power, and compute efficiency.
• Adapt and integrate optimized models to inference runtimes such as TensorRT, ONNX, TFLite, SNPE, OpenVINO, or TVM.
• Implement quantization, pruning, and other model compression techniques.
• Support for integration of AI workloads into embedded software stacks running on Linux, RTOS, or bare-metal systems.
• Participate in the team's software processes to ensure code quality & maintenance, including -- requirements and design documentation, test-plan generation and execution, peer design and code reviews
• Stay current with advancements in edge computing, AI inference frameworks, and compiler toolchains.

Required skills / Whom we are looking for:

• Bachelors or Masters degree in related engineering field with 4 to 8 years of hands-on experience in experience in embedded AI/ML development, with a focus on model optimization and deployment.
• Proficiency in in C/C++ and Python programming, Intrinsic or Assembly based optimization methods using instruction pipeline and latency optimal designs, Modular and Object-Oriented programming skills
• Experience working with heterogeneous computing platforms (e.g., CPU + GPU/DSP/NPU). Must have exposure and development experience on one or more DSPs/NPUs for example ARM-NEON, TI C6x/C7x DSP, Tensilica Vision DSPs, CEVA DSPs, Qualcomm Hexagon HVX DSP
• In-depth knowledge Processor/SoC architecture – VLIW and SIMD, DMA, cache, memory architecture etc.,
• Working experience in machine learning technologies such as CNN, transformers, quantization algorithms and approaches on embedded systems
• Hands-on experience with any of the AI frameworks such as TensorFlow, PyTorch, or ONNX and familiarity with inference toolkits such as TensorRT, SNPE, TFLite, TVM, or OpenVINO.
• Familiarity with build systems (e.g. make, cmake, GCC, Eclipse, Visual Studio, ARM Development Tools)
• Familiarity with debugging tools such as GDB, JTAG, and performance profiling tools.
• Well verse with software development life cycle and efficient use of associated tools like Git, SVN, JIRA etc.,
• Strong communication skills and the ability to work effectively in a collaborative, cross-functional team environment.
• Detail-oriented with a focus on delivering high-quality, reliable software.
• Self-motivated with a strong passion for embedded AI systems and technology.