Artificial Intelligence Researcher

About the Company We are a stealth-stage AI systems company backed by successful second-time founders. We are building next-generation infrastructure designed to tackle complex performance, compute, and scalability bottlenecks in the AI stack.Driven by a high-bar engineering culture, we are solving deep systems challenges to enable high-performance execution for modern AI workloads. If you thrive on deep technical problems, performance engineering, and high-impact systems, we’d love to talk.About the Role & TeamWe are looking for exceptional AI Researchers to join our core research group. In this role, you will work across the entire stack: from core model architecture and training algorithms, down to kernel optimization and serving engines, and up to the application harness.This is a hands on research role: while you will engage in deep algorithmic research and state of the art innovation, you will also implement models, write code, and see your research directly applied to solve high impact, real world problems.What We DoWe specialize in optimizing large language model (LLM) inference and performance at scale. Our technology bridges the gap between raw model capacity and production grade efficiency, maximizing latency, throughput, cost efficiency, and output quality across complex AI workflows.Key Responsibilities & Focus AreasDepending on your expertise, your focus will span one or more of the following areas:Training & Optimization: Mid training and post training of large language models at scale (multi node distributed training).Alignment & Model Quality: Distillation, RL/RLHF, alignment, guardrail management, and debiasing.Efficient Inference: Model pruning, quantization, speculative decoding algorithms, and guided decoding for structured outputs/tool calls.Systems & Engines: CUDA kernel optimization, internal modifications to LLM serving engines (e.g., vLLM), KV cache aware routing algorithms, and offloading strategies.Agent Harnessing: Profiling, optimizing agentic execution loops, smart LLM routing, and call optimization.Requirements & QualificationsEducation: MSc or PhD in Computer Science, Machine Learning, Mathematics, or a related quantitative field (ongoing studies or candidates in progress are welcome).Research Background: Strong background in at least one of the focus areas listed above. A track record of recent publications in top tier conferences or journals (e.g., NeurIPS, ICML, ICLR, ACL, EMNLP) or experience in a renowned university research lab is a strong advantage.Hands on Mindset: Proficiency in Python and deep learning frameworks (PyTorch, JAX, etc.). You enjoy writing code, prototyping algorithms, and implementing your own ideas rather than handing them off.Open Scope & Seniority: We are open to various levels of seniority, from talented students to senior or principal researchers.