---
title: 'Software Engineer – GPU Kernel at FriendliAI'
canonical: 'https://feeny.ai/job/software-engineer-gpu-kernel-friendliai-seoul-yg1eq3ks7m5a'
type: 'job'
last_seen: '2026-09-05'
---

# Software Engineer – GPU Kernel at FriendliAI

- **Company:** FriendliAI
- **Location:** Seoul, South Korea
- **Employment:** full-time
- **Work type:** onsite
- **Posted:** 2026-04-22
- **Last confirmed live:** 2026-09-05
- **Apply:** https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735

## Job description

## ABOUT THE JOB

FriendliAI is looking for a GPU Kernel Engineer to design, build, and optimize the low-level compute kernels that power our large-scale, GPU-accelerated AI inference platform. You will be delivering world-class inference speed across NVIDIA and AMD GPUs. With our recent $20M funding, we are scaling our team to meet market demand.

This is a deeply technical, high-impact role where you will write GPU code, implement advanced optimizations. As part of our engine team, you will contribute directly to the company’s proprietary inference engine which supports over 600,000 models on Hugging Face. You will work with the inventors of continuous batching and collaborate with the platform team to deploy your work into production.

## KEY RESPONSIBILITIES

- Design, implement, and optimize high-performance GPU kernels for AI inference (e.g., GEMM, attention, routing)
- Develop and maintain GPU code in CUDA and C++, including low-level assembly when needed
- Implement reduced-precision and quantized kernels (FP8/FP4) for low-latency or high-throughput inference
- Benchmark and ensure cross-vendor performance parity between NVIDIA and AMD hardware
- Contribute to internal GPU libraries and tune performance of performance-critical components
- Accelerate multi-modal model pipelines
- Investigate and integrate next-generation GPU features

## QUALIFICATIONS

- 3+ years of experience in GPU programming, HPC, or performance-critical systems
- Bachelor’s or Master’s degrees in Computer Science, Computer Engineering, Electrical Engineering, or a related field
- Strong proficiency in CUDA for NVIDIA GPUs or ROCm/HIP for AMD GPUs
- Deep understanding of GPU architecture: warps, threads, memory hierarchy, synchronization, and latency-throughput trade-offs
- Proficiency in C++
- Experience with GPU profiling and performance tuning
- Strong numerical background with understanding of precision trade-offs and quantization techniques

## PREFERRED EXPERIENCE

- Experience optimizing transformer, multi-modal, or Mixture-of-Experts (MoE) architectures at the kernel level
- Familiarity with the latest GPU libraries and frameworks (CUTLASS, Triton, …)
- Inter-GPU communication programming experience
- Open-source contributions related to GPU performance or ML acceleration
- Research or conference presentations on GPU optimization, HPC, or numerical computing

## BENEFITS

- Flexible working hours
- Daily lunch and dinner provided; unlimited snacks and beverages
- Supportive and highly collaborative work environment
- Health check-up support and top-tier equipment/hardware support
- A front-row seat to the generative AI infrastructure revolution
- Competitive compensation, startup equity, health insurance, and other benefits.

## ABOUT FRIENDLIAI

FriendliAI is building the world’s best AI inference platform that makes large language and multi-modal models fast, efficient, and deployable at scale. We power high-throughput, low-latency AI workloads for organizations worldwide and integrate directly with Hugging Face, giving developers instant access to over 600,000 open-source models.

We are a small, fast-moving team doing work that matters at one of the most exciting moments in the history of technology. With our world-class inference engine, we are building a platform that the AI industry can actually rely on.

## About FriendliAI

## Company Overview
- **One-liner**: FriendliAI is The Frontier AI Inference Cloud, providing a highly optimized platform for deploying, scaling, and monitoring large language and multimodal models with unmatched speed and cost efficiency.
- **Entity Type**: Private (Seed stage; $26.7M total funding)
- **Headquarters**: San Francisco, California, United States (with offices in Seoul, South Korea and Redwood City, CA)
- **Founded**: 2021
- **Founders**: Byung-Gon Chun (CEO) and Gyeong-In Yu (CTO) – the researchers who invented the continuous batching technique now standard in AI inference.

## Core Business
- **Primary industry**: AI Infrastructure / Cloud Inference Platform (Infrastructure as a Service)
- **Target customers**: B2B – enterprises and AI teams deploying generative AI and agent workloads at production scale; focuses on open-weight and custom models.
- **Mission**: Democratize access to production-grade AI so teams can focus on building great products.

## Products & Services
- **[FriendliAI Inference Cloud](https://friendli.ai/)**: A fully managed platform that instantly deploys 580,000+ Hugging Face models (language, audio, vision) with one click. Supports bring-your-own fine-tuned or proprietary models. Key features:
  - Custom GPU kernels, smart caching, continuous batching, speculative decoding, and parallel inference.
  - 2×+ faster inference than standard solutions, up to 3× faster than vLLM.
  - 50% to 90% cost savings relative to closed model APIs.
  - 99.99% uptime SLAs, geo-distributed infrastructure, enterprise-grade fault tolerance.
  - SOC 2 Type II and HIPAA compliant (March 2026).

## Market Standing
- **Valuation/Market Cap**: Not disclosed (private company)
- **Key Metric**: Total funding $26.7M (two seed rounds: $6.7M in 2021 and $20M led by Capstone Partners in September 2025)
- **Notable Investors/Partners**: Capstone Partners (lead investor in latest round), plus two other investors in the first seed round.
- **Growth Signals**:
  - Headcount grew 90% YoY to ~45 employees (as of mid-2026).
  - Active job postings: 19 positions (up 1800% YoY) – strong hiring momentum.
  - Appointed Brian Yoo (former Moloco COO) as Chief Business Officer in April 2026 to drive hypergrowth.
  - Achieved SOC 2 Type II and HIPAA compliance in March 2026, opening regulated markets.
  - LinkedIn followers grew 257% yearly (8,221 followers).

## Competitive Advantages
- **Inventor of continuous batching** – the technique is now industry standard, giving FriendliAI deep technical expertise.
- **Purpose-built inference engine** that constantly evolves for state-of-the-art models, delivering 2–3× speed improvements over alternatives like vLLM.
- **Massive model library** – 580,000+ Hugging Face models deployable instantly with zero manual optimization.
- **Enterprise-grade reliability** – 99.99% SLA, SOC 2 Type II, HIPAA, multi-cloud scaling.
- **Cost leadership** – 50–90% savings vs. closed model APIs, maximizing tokens per dollar.

## Strategic Focus
- **Hypergrowth** – scaling sales, engineering, and solutions architecture teams globally (US and Korea).
- **Compliance and enterprise readiness** – recent SOC 2/HIPAA certification to serve healthcare and regulated industries.
- **Multi-cloud and geo-distribution** – expanding infrastructure footprint for global low-latency inference.
- **Open-weight and custom model support** – enabling model ownership and flexibility for enterprises.

## Why Work Here
- **Culture**: “Boldest innovations come from great teams” – passionate, humble, curious builders. Mission-driven to democratize AI.
- **Work policy**: Hybrid (San Francisco and Seoul offices); some roles are in-office (Seoul) or hybrid (SF). Employees engage in a mix of remote and on-site work.
- **Engineering focus**: Heavy emphasis on AI inference engine, GPU kernels, Python developer tools, AI agents, backend, and platform security. High technical bar.
- **Perks**: Not explicitly listed, but fast-growing startup with significant impact in the AI infrastructure space; opportunity to work on cutting-edge inference optimization.
- **Locations**: San Francisco (HQ), Redwood City, CA, and Gangnam-gu, Seoul – global team with cross-cultural collaboration.

## Sources
1. [friendli.ai](https://friendli.ai/)
2. [friendli.ai/careers](https://friendli.ai/careers)
3. [jobs.ashbyhq.com/friendliai](https://jobs.ashbyhq.com/friendliai)
4. [builtin.com/company/friendliai](https://builtin.com/company/friendliai)
5. [linkedin.com/company/friendliai](https://www.linkedin.com/company/friendliai)

## Other roles at FriendliAI

- [Director of Product Management](https://feeny.ai/job/director-of-product-management-friendliai-san-francisco-vckbhh439mws) — San Francisco, CA
- [Software Engineer - Full Stack](https://feeny.ai/job/software-engineer-full-stack-friendliai-seoul-2kvnfm2hejx8) — Seoul, South Korea
- [Software Engineer – Cloud Infrastructure](https://feeny.ai/job/software-engineer-cloud-infrastructure-friendliai-san-francisco-z8fjtpn12kpn) — San Francisco, CA
- [Software Engineer - Cloud Infrastructure](https://feeny.ai/job/software-engineer-cloud-infrastructure-friendliai-seoul-rf473fpa9850) — Seoul, South Korea
- [Account Executive](https://feeny.ai/job/account-executive-friendliai-san-francisco-z6e1g3bcymfg) — San Francisco, CA
- [Software Engineer – Python Developer Tools](https://feeny.ai/job/software-engineer-python-developer-tools-friendliai-seoul-z74wd9kp8fex) — Seoul, South Korea
- [Software Engineer – AI Inference Engine](https://feeny.ai/job/software-engineer-ai-inference-engine-friendliai-seoul-8pdpe18z4ejk) — Seoul, South Korea
- [Customer Success Engineer (contract based)](https://feeny.ai/job/customer-success-engineer-contract-based-friendliai-seoul-9ang4dybr9c9) — Seoul, South Korea
- [Software Engineer - Senior Backend](https://feeny.ai/job/software-engineer-senior-backend-friendliai-san-francisco-h5x2ghcnfrg7) — San Francisco, CA
- [Software Engineer – AI Agents](https://feeny.ai/job/software-engineer-ai-agents-friendliai-san-francisco-7v0cdvt2ded6) — San Francisco, CA