AI Researcher, Agent

At Goaly, our mission is to make custom AI affordable for every business. Our founding team comes from the front lines of top AI labs and tech giants (Meta MSL, TikTok AI, Google DeepMind, xAI, Microsoft Research, etc.), where we built large-scale training infrastructure powering trillion-parameter models and scaled GenAI models to a global user base. Now, we are building something we wish we had before: a platform that makes training and adapting custom AI affordable for all modern companies, not just Big Tech. Our north star is ambitious: for a domain‑specific task, reach 90% of SOTA performance at less than 10% of the cost. To get a taste of what we are doing, see our first tech blog . Role overview As an AI Researcher, Agent, you will own the full research lifecycle for agentic AI: from training agent‑specialized models to building the orchestration stacks and simulated environments that make large‑scale agentic RL possible. You will design and run experiments that redefine what autonomous AI systems can do—and you will publish the results that move the field forward. Core responsibilities Train agent‑specialized models: Research and develop training recipes—including SFT, RLHF, and agentic RL—that produce models capable of robust multi‑step reasoning, tool use, and long‑horizon task completion. Build agent orchestration: Design and implement scalable orchestration frameworks for multi‑agent systems, including planning, memory, tool integration, and inter‑agent communication protocols. Simulate agentic environments: Create and maintain high‑fidelity simulated environments that support diverse task distributions, controllable difficulty, and fast data generation for agentic RL training at scale. Design domain‑tailored eval benchmarks: Build evaluation frameworks that capture real‑world performance gaps—covering instruction following, tool‑use accuracy, error recovery, and long‑horizon goal completion across targeted domains. Advance research and publish: Translate experimental results into publications at top venues (NeurIPS, ICML, ICLR) with full internal support for compute, datasets, and writing. Requirements Ph.D. or Master’s degree in CS, AI, Math, or related fields (exceptional industry researchers are welcome). Deep understanding of Transformer architectures, RLHF, and large‑scale LLM training from data curation through post‑training optimization. Hands‑on experience building or researching agentic systems—tool‑use, ReAct/chain‑of‑thought, multi‑step planning, or multi‑agent coordination. Strong PyTorch or JAX skills paired with clean engineering practices and reproducible experiment design. Experience designing rigorous evaluation benchmarks or simulation environments for AI systems. Adaptability: a fast learner with a strong sense of ownership, capable of working across research, systems, and product simultaneously. Bonus Points First‑author publications at top conferences (NeurIPS, ICML, ICLR, CVPR, ACL) on topics such as RLHF, agentic reasoning, tool use, or multi‑agent systems. Experience with large‑scale agentic RL pipelines (e.g., asynchronous training, reward shaping, multi‑agent environments). Contributions to open‑source agent frameworks or evaluation harnesses (e.g., AgentBench, WebArena, SWE‑bench, ToolBench). High‑star open‑source projects on Hugging Face or competitive ML achievements (Kaggle medals, competition wins). Why Join Us? Expert Mentorship: Partner with AI veterans who have trained trillion‑parameter models at scale and applied them to solve real‑world problems in billion‑user products. Compute Freedom: Access to abundant GPU cluster resources—don’t let your creativity be limited by compute. Flat Culture: Flat management structure that rejects office politics and values only technology and results. Competitive Compensation: Competitive full‑time offers, huge upside, and extra equity incentives when hitting key milestones. #J-18808-Ljbffr Goaly