Sr. AI Inference Software Engineer Compiler
Website Flex Logix
Flex Logix is the leading provider of reconfigurable computing technology for both AI inference and eFPGA IP solutions. Our offerings push the leading edge of hardware, software and system design; pioneering new approaches to important problems.
- Our InferX X1 is the industry’s most-efficient AI edge inference accelerator that brings AI to the masses in high-volume applications by providing a new silicon efficient dynamic logic paradigm for inference processing. InferX achieves GPU-level inference performance with a fraction of the die area and memory footprint.
- Our EFLX embedded FPGA (eFPGA) IP enables any SOC design to flexibly handle changing protocols, standards, algorithms, and customer requirements and enables reconfigurable accelerators that speeds key workloads up to 1000x compared to a general purpose processor. EFLX eFPGA is available in a wide range of process technologies and supports designs ranging from low cost microcontrollers to 5G baseband processing solutions.
Flex-Logix Inference Software team is looking for Inference Software Compiler Engineers to be a part of its excellent team responsible for the Infer-X Model Compiler. The compiler generates Verilog Code for the Flex-Logix eFPGA platform which controls the nnMAX/TPU computation blocks and memory connections. This is an exceptional opportunity to develop the technology that breathes life into AI inferencing solutions targeting systems in medical, industrial, automotive and other Enterprise edge applications.
In this role, you will be responsible for developing highly optimized libraries fundamental to the operation of the AI accelerator. You’ll be working with multiple software teams, deep learning scientists, and hardware designers, to implement the features needed to accelerate the next generation of machine learning algorithms.
If you’re a collaborative engineer or scientist who has a passion for innovation, solving challenging technical problems and doing impactful work you need to join the team at Flex Logix!
Responsibilities
- Collaborate with the applications, softlogic, architecture, and compiler teams to determine low-latency, high-throughput schemes for executing neural networks on our ML accelerator.
- Understand low level operator/algorithm implementation and develop APIs to invoke individual operators at the graph level.
- Make minor modifications to operator/algorithm implementations.
- Devise unit tests for new deep learning models.
- Evaluate the performance of the neural network on silicon.
- Design and develop supporting libraries that run some neural network operators on CPUs.
- Diagnose and fix performance and integration issues across the software stack using simulators and hardware.
Requirements
- BS or higher in computer/software engineering, electrical engineering, or related field.
- 2+ years commercial development or equivalent hands-on experience.
- Highly proficient in C/C++.
- Strong mastery of data structures and algorithms, especially graph data structures.
- Experience developing and/or analyzing algorithms.
Preferred Skills
- Experience with modern C++ (e.g. C++17).
- Ability to prototype complex systems with scripting languages such as Python.
- An ability to learn rapidly in order to solve complex problems.
- Proficient at using objective oriented programming (OOP) design principles.
- Experience with design or verification of digital systems using verilog or RTL.
- Experience writing applications for SIMD processors or accelerators like GPUs or FPGAs.
- Strong understanding of computer architecture.
- Familiarity with analyzing machine learning networks using a Deep Learning framework (e.g. Pytorch, Tensorflow, etc.).
We are looking for passionate team members, to be part of an aggressive, venture-backed startup team that is changing chip architecture. Must be entrepreneurial, innovative problem solver, willing to work hard and have fun.
As we continue to grow and expand our company, we are hiring for all office locations. You must live near one of our main offices located in: Mountain View (CA), Austin (TX), Chicago (IL) or Vancouver (BC). We offer a flexible work schedule.
You must have US citizenship or permanent residency (“green card”) or hold a current H1-B visa to work in United States.
Flex Logix recruits, employs, trains, compensates and promotes regardless of race, religion, color, national origin, sex, disability, age, veteran status, and other protected status as required by applicable law.
The Data Crisis is Unfolding – Are We Ready?