Find your next opportunity
Browse fresh openings from leading companies and apply with a resume tailored to the role.
Browse expert rolesBrowse expert rolesBrowse fresh openings from leading companies and apply with a resume tailored to the role.
Browse expert roles14200 openings
5713-5720 of 14200Kimchi
Free to start · No card · 5 credits the moment you sign up
Great models are raw ingredients. Kimchi is what happens when you ferment them with good engineering. We started inside CAST AI helping companies run LLMs on their own Kubernetes clusters; today we run our own coding harness on any model you pick, and we run the models underneath it ourselves.
Kimchi Inference
One API line and a customer's existing client just works: DeepSeek, GLM, Kimi, MiniMax, Nemotron served serverless from our own GPU fleet (OpenAI- and Anthropic-compatible, pay per token, spend visible per user, team, and key). When control matters, the same models run on the customer's own cluster instead. Behind that one line sits the machinery you'll own: SGLang and vLLM engines, B300 nodes, a shared cluster filesystem that holds every checkpoint exactly once, KV caches riding coding traffic with heavy prefix reuse, and the benchmark harness that gates every change we ship.
Made by Cast AI — a Miami-headquartered, globally distributed team across 34 countries that hit unicorn status in January 2026 with a $50B+ backed investment pushing its valuation past $1 billion.
Throughput. Latency. Cold starts. Cost per token.
Move those four numbers in the right direction and two things happen: customers get faster, cheaper inference, and our margins improve. That's the entire thesis of this role. Every loader you tune, every scheduler decision you land, every kernel you stare at shows up in a customer's p99 and on our P&L. Your mandate is the inference stack's optimization, whatever it takes, and the technical direction of how we optimize it is yours to argue with data.
The problem: the checkpoints run 0.3-1.5 TB, and the "right" serving configuration depends on traffic shape, sequence-length distribution, batch dynamics, GPU SKU, memory bandwidth, and quantization tolerance, all variables that shift week to week. Most teams pick a configuration once, over-provision GPUs, and absorb the cost. We don't: we took Kimi-K3 (1.5 TB) from over an hour cold-to-serving to under twelve minutes with a shared filesystem, O_DIRECT weight loaders, and node-local JIT caches, and we're nowhere near done. Prefill/decode disaggregation and tiered KV caches are next.
Stack: Python; SGLang; vLLM; TensorRT-LLM; PyTorch; fastsafetensors; CUDA-adjacent tooling; Kubernetes; virtiofs shared filesystems; ClickHouse; OpenTelemetry; Prometheus; Grafana; GitLab CI.
As part of our standard hiring process, we would like to inform you that a background check may be conducted at the final stage of recruitment through our third-party provider, Checkr.
Please note that Cast AI does not provide any form of visa sponsorship/work permit.
#LI-Remote
Start with a role or skill you want to use in your next job. Use the location and role filters to narrow the results, then open a listing to review the description. A title alone rarely explains the level of responsibility, working arrangement, or experience a team needs. Compare those details before deciding which opportunities deserve a tailored application.
Employers use different titles for similar work. If an exact phrase produces no results, try a shorter title, a core skill, or an alternative name for the role. Remove one filter at a time to understand which requirement is narrowing your search. A broader search can reveal suitable work without asking you to change your actual qualifications.
For each promising listing, read the responsibilities and separate essential requirements from preferences. Note the location, schedule, employment type, and compensation when provided. Missing details are questions to confirm with the employer. A remote opening may still require a particular country, work authorization, or time-zone overlap, so check the full posting before treating it as a location-independent opportunity.
Choose a few examples that demonstrate the skills this employer names. Explain your contribution and its outcome, using accurate numbers where available. Put the most relevant experience first and keep your resume easy to scan. Avoid adding technologies, qualifications, or achievements simply because they appear in a description. Your application should make your existing strengths easier to understand.
Follow the employer’s instructions for resumes, cover letters, work samples, and screening questions. Review the exported file before submitting it: confirm that contact details are correct, text can be selected, and links reach the intended pages. A resume check can help you spot unclear writing before you apply. Its feedback is advisory and does not predict a hiring decision.
Listings are collected from external sources and may change after they appear here. The employer or recruiting destination controls whether applications are open. If an original link is unavailable, check the company’s careers page for the title or move to another opening. An employer website link is a place to check current vacancies, rather than confirmation that a particular job is still accepting applications.
Keep a simple record of the company, role, job reference, application date, and documents you sent. Record any follow-up instructions and the next action you plan to take. This makes it easier to prepare for interviews, compare opportunities, and avoid applying repeatedly to the same opening through different boards. Review your shortlist as employers update their listings and prioritize opportunities with clear requirements that match the work you want to do.