Principal Software Engineer - Compilers

Who We Are

This role is part of theKinara Group at NXP Semiconductors, combining the agility and innovation of a startup with the scale, stability, and global reach of a leading semiconductor company.

Kinara was founded onpatented technology developed during the founders’ Ph.D. research at Stanford University, with a mission to redefine AI inference at the edge through tightly integrated hardware and software. Following Kinara’s acquisition, this mission continues within NXP, with the Kinara team operating as adedicated AI accelerator and software group focused on pushing the boundaries of edge AI.

The team has deliveredtwo generations of AI silicon shipping in volume, supported by amature and production‑ready software stack, making it one of the most proven edge AI platforms in the industry. Kinara technology is used bymarquee customers, including GAFA‑class companies, a leading global e‑commerce provider, and Tier‑1 PC OEMs.

As part of NXP, the Kinara Group now benefits fromworld‑class silicon execution, manufacturing scale, and customer access, while retaining its startup culture, deep AI expertise, and compiler‑driven engineering mindset. Joining this team offers thebest of both worlds: the opportunity to work on cutting‑edge AI compiler and silicon technology, with real‑world impact and long‑term stability.

What We Do

We buildgame‑changing edge AI solutions that enable faster, smarter, and more power‑efficient decision making where every millisecond matters.

At the core of our platform areAra AI inference processors, purpose‑built to deliverunrivaled deep learning performance per watt at the edge. Paired with ourcomprehensive SDK and software toolchain, these processors enable developers to deploy and optimize complex AI models for real‑time inference under tight latency and power constraints.

Kinara’s hardware–software co‑design approach allows high‑performance AI to beseamlessly embedded into edge devices, enabling applications across vision, industrial automation, consumer electronics, and intelligent systems. Our solutions focus onperformance, efficiency, and scalability, helping customers bring advanced AI capabilities out of the cloud and into the real world.

Edge AI is at an inflection point, and the Kinara Group within NXP is uniquely positioned to play aleading role in its growth, delivering production‑proven silicon, a robust compiler and SDK stack, and a clear roadmap for next‑generation intelligent edge devices.

Role Overview

We are looking for an experiencedFront-End Compiler Engineer with 10+ years of experience to lead the design and development of ourAI/ML compiler front-end. The role focuses on building scalablemodel conversion pipelines that translate models fromONNX, TensorFlow, and PyTorch into our internalIntermediate Representation (IR).

The ideal candidate will have deep expertise incompiler design,graph-based optimizations,machine learning frameworks, and modern AI architectures includingTransformers andLarge Language Models (LLMs). You will collaborate with compiler backend, runtime, performance, and hardware teams to deliver a high-performance and production-ready AI compiler stack.

Key Responsibilities

• Design and developcompiler front-end architecture andmodel conversion frameworks.
• Build and maintainPython-based converters forONNX,TensorFlow, andPyTorch.
• Implementgraph construction,graph transformation,IR lowering, and optimization pipelines.
• Developgraph optimization passes includingoperator fusion,canonicalization,simplification, anddecomposition.
• Design scalablepattern-matching andgraph-rewrite frameworks.
• Enable support for modern AI architectures such asCNNs,Transformers, andLLMs.
• Drive validation, testing infrastructure, regression testing, and CI/CD integration.
• Debug and resolve issues across model conversion, optimization, and IR generation stages.
• Providetechnical leadership, mentor engineers, and collaborate across teams.

Required Skills & Experience

• 10+ years of experience incompiler development,ML systems, orAI infrastructure.
• StrongPython programming and software architecture skills.
• Deep understanding of:
  • Intermediate Representations (IRs)
  • Graph-based computation models
  • Compiler transformations and optimizations
  • Pattern matching and graph rewriting
• Hands-on experience withONNX,TensorFlow, andPyTorch.
• Experience ingraph parsing,graph transformations, andML model optimizations.
• Strong understanding ofCNNs,Transformers, andLLM architectures.
• Excellentdebugging,problem-solving, andtechnical leadership skills.

Good to Have

• Experience withMLIR,TOSA,StableHLO, orTorch-MLIR.
• Familiarity with AI compiler stacks such asIREE,TVM,XLA,TensorRT, orONNX Runtime.
• Experience targetingGPUs,NPUs,DSPs, or custom AI accelerators.
• Knowledge ofquantization,model optimization, and deployment workflows.

More information about NXP in India...