Edge AI Optimization Engineer

Edge AI Optimization Engineer – Compiler & Deployment Enablement

Position Overview

A global semiconductor and embedded systems organization is seeking an experienced AI/ML Compiler Engineer to support the development of a unified edge AI deployment platform.

This role plays a central part in building a scalable, MLIR-based compiler infrastructure that enables consistent model ingestion, optimization, and code generation across a wide range of embedded and heterogeneous hardware targets.

The position operates at the intersection of artificial intelligence systems, compiler technology, and hardware enablement, and includes significant technical leadership responsibilities.

Role Purpose & Mission

The mission of this role is to design, develop, and advance a modern compiler and deployment framework that:

• Serves as the primary front-end and optimization pipeline for edge AI workloads.

• Enables scalable “bring-your-own-model” deployment across diverse hardware platforms.

• Supports unified intermediate representations and modern compilation workflows.

• Facilitates integration into internal AI software ecosystems and product platforms.

Key Responsibilities

1. Front-End and IR Pipeline Development (MLIR-Based)

• Design, implement, and maintain model ingestion pipelines for major AI frameworks, including PyTorch, ONNX, TensorFlow, and TFLite.

• Develop and extend MLIR dialects to interface with proprietary and third-party hardware IP.

• Create specialized pipelines for quantized models and large-scale neural networks.

2. Model Optimization and Hardware Lowering

• Design and implement optimization passes, IR transformations, and backend lowering flows.

• Enable efficient code generation and operator support for heterogeneous hardware targets.

• Support device-level performance tuning and microcode generation where applicable.

3. Exploratory and Advanced Compiler Research

• Contribute to innovation tracks such as:

  • Training graph ingestion for on-device learning

  • Automated performance tuning

  • Heterogeneous scheduling and workload partitioning

  • Unified runtime binding and API layers

• Evaluate emerging compilation and runtime technologies to guide long-term platform evolution.

4. Integration into Software Platforms

• Deliver compiler components for internal SDKs, runtimes, and deployment environments.

• Develop standardized model packaging and execution formats.

• Support integration with automotive, industrial, and embedded software stacks.

5. Collaboration and Knowledge Transfer

• Work with internal product and engineering teams to transition research outputs into production systems.

• Collaborate with external partners and open-source communities to accelerate innovation.

• Contribute to technical disclosures, internal standards, and best-practice documentation.

Required Qualifications

• Strong background in compiler design and implementation, including experience with MLIR, LLVM, or modern ML compilers (e.g., TVM, IREE, XLA).

• Deep understanding of AI/ML models, quantization pipelines, and hardware-aware optimization techniques.

• Proficiency in C++ and Python, with experience developing compiler passes, toolchains, or device backends.

• Practical experience with model export and conversion flows from frameworks such as PyTorch, ONNX, and TensorFlow.

• Experience working with embedded and heterogeneous computing platforms (CPUs, NPUs, DSPs, accelerators).

Preferred Qualifications

• Hands-on experience with MLIR dialect development and IR transformation pipelines.

• Experience with ahead-of-time compilation and runtime integration workflows.

• Contributions to open-source compiler or machine learning infrastructure projects.

• Knowledge of automotive or industrial system constraints, including safety, determinism, and real-time performance requirements.

Personal Attributes

• Strong technical leadership and systems-thinking mindset.

• Ability to work across hardware, software, and research domains.

• High level of initiative in exploratory and long-term technical development.

• Structured, documentation-oriented working style.

• Effective communication skills for multidisciplinary collaboration.