--- title: 'Member of Technical Staff - Foundations at Tzafon' canonical: 'https://feeny.ai/job/member-of-technical-staff-foundations-tzafon-san-francisco-y8me0a5e3at9' type: 'job' last_seen: '2026-09-16' --- # Member of Technical Staff - Foundations at Tzafon - **Company:** Tzafon - **Location:** San Francisco, CA / Tel Aviv, Israel / Zurich, Switzerland - **Employment:** full-time - **Work type:** hybrid - **Posted:** 2025-10-10 - **Last confirmed live:** 2026-09-16 - **Apply:** https://jobs.ashbyhq.com/tzafon/d24d8aa3-099e-47ec-82b8-22c1649727ce ## Job description Tzafon is a foundation model lab building scalable compute systems and advancing machine intelligence, with offices in San Francisco, Zurich & Tel Aviv. We’ve raised over $12m in funding to advance our mission of expanding the frontiers of machine intelligence. We're a team of engineers and scientists with deep backgrounds in ML infrastructure & research. Founded by IOI and IMO medalists, PhDs, and alumni from leading tech companies, such as Google Deepmind, Character, and NVIDIA, we train models and build infrastructure for swarms of agents to automate work across real-world environments. You'll work between our product and post-training teams to ship Large Action Models that actually work. Build evals, benchmarks, and fine-tuning pipelines. Define what good model behavior means and make it happen at scale. ## What you'll do - Design and execute large scale training runs on our clusters - Build and optimize distributed training infrastructure across massive multi-node systems - Implement post-training pipelines at scale - Develop data pipelines that process and filter trillions of tokens for pre-training - Research and implement architectural improvements, scaling laws, and training optimizations - Debug training instabilities, loss spikes, and convergence issues in long-running jobs - Build tooling for cluster utilization, fault tolerance, and checkpoint management - Write custom CUDA/Triton kernels to optimize critical training operations (attention, normalization, activations) - Collaborate on research that advances the state of the art in foundation model training We're looking for - Deep experience pre-training or post-training foundation models on large clusters - Expert-level at Python and ML frameworks (PyTorch, JAX, Torchtitan) - Strong systems skills: distributed training, FSDP/ZeRO, tensor parallelism, pipeline parallelism - Experience writing performant CUDA or Triton kernels for ML workloads - Track record of running stable multi-week training jobs and debugging distributed training failures - Understanding of cluster scheduling, networking bottlenecks, and GPU/TPU performance optimization ## Preferred Experience - Trained foundation models at major AI labs (OpenAI, Anthropic, Google DeepMind, Meta, xAI, etc.) - Worked on large scale RL runs - Optimized critical training kernels (FlashAttention, fused optimizers, custom kernels) - Published research at top ML conferences (NeurIPS, ICML, ICLR) - Contributions to open source ML infrastructure (PyTorch, JAX, vLLM, etc.) - Experience with training data pipelines, data quality research, or synthetic data generation Life at Tzafon - Full medical, dental, and vision coverage, plus 401(k) in the us - Office in SF, Zurich, and Tel Aviv - Early-stage equity in a future-defining company Visa sponsorship: We do sponsor visas! However, we aren't able to successfully sponsor visas for every role and every candidate. But if we make you an offer, we will make every reasonable effort to get you a visa, and we retain an immigration lawyer to help with this. Compensation starts at $200k-$500k + equity package, depending on experience & location. We also offer a referral bonus of $5k for referral of successful hires (send to careers@tzafon.ai). ## About Tzafon ## Company Overview - **One-liner**: Tzafon is an applied AI research lab building scalable compute systems and foundation models for autonomous agents that interact with digital and physical environments. - **Entity Type**: Private (Seed stage – $16.7M total funding across multiple rounds) - **Headquarters**: San Francisco, California, United States / Tel Aviv, Israel (dual headquarters); also has offices in Rochester, NY and Zurich, Switzerland - **Founded**: 2024 - **Founders**: Noah Löfquist (CEO) and unnamed co-founders ## Core Business - **Primary industries**: Technology, Information and Internet; Artificial Intelligence; Research - **Target customers**: B2B (enterprises requiring automation) and B2C (individuals using autonomous assistants); also serves developers and AI researchers - **Mission / Purpose**: “Advancing machine intelligence” by building agents that operate across any digital (and eventually physical) environment, with the belief that AGI remains the most important unsolved problem this century ## Products & Services - **Lightcone**: An autonomous AI agent designed to operate seamlessly on a user’s behalf across any app or platform, reducing friction between intention and action. - **Foundation Models (computer-use models)**: Trained for real-world interaction and web-based tasks; the first model achieved top results in OSWorld benchmarks. - **Multi-agent Systems**: Advanced frameworks that enable agentic AI to collaborate across digital environments, leveraging reinforcement learning and proprietary architectures. - **Compute Infrastructure**: Scalable systems built to support large-scale model training and inference. ## Market Standing - **Valuation / Market Cap**: Not publicly disclosed - **Key Metric**: Total funding of **$16.7M** (pre-seed/seed rounds from 2025); Annual revenue ~$193K (early stage) - **Notable Investors / Partners**: HV Capital (lead), Streamlined VC, Kakao VC, Oliver Jung; angel investors from OpenAI, xAI; advised by Neej Parikh and Michal Valko - **Growth Signals**: - Headcount grew 137.5% YoY (from ~7 to 15–18 employees) - Top performance on OSWorld web-based benchmarks - Opened offices in Tel Aviv, Zurich, and San Francisco - Launching Lightcone publicly as a user-facing product ## Competitive Advantages - **Proprietary multi-agent architecture** enabling cross-platform autonomy (digital and eventually physical). - **Team composition** from top AI labs (DeepMind, Google, Decart) and quantitative firms (Jane Street, Citadel), plus elite units (Unit 81, Unit 8200). - **Early demonstration of state-of-the-art results** on web-agent tasks (OSWorld). - **Dual HQ model** (SF + Tel Aviv) attracting global talent with domain expertise. ## Strategic Focus - **Scaling compute infrastructure** to support larger, smarter models and increased demand. - **Launching and maturing Lightcone** as a seamless digital assistant for end users and enterprises. - **Advancing agentic AI capabilities** (reinforcement learning, continuous learning, multimodality) to enable real-world interaction beyond the screen. ## Why Work Here - **Culture**: Attracts “curious minds from a wide range of disciplines”; emphasis on global perspectives and respect for diverse experiences. - **Work policy**: Hybrid – employees combine remote work with on-site time at one of two offices in the US (San Francisco, Rochester) or in Tel Aviv / Zurich. - **Team & growth**: Small but rapidly growing team (~18 people) with high ownership; flat structure (5 people at C-level). - **Impact**: Directly working on foundational AGI problems with the possibility to shape product direction from an early stage. ## Sources 1. [tzafon.ai](https://www.tzafon.ai) 2. [LinkedIn – Tzafon](https://www.linkedin.com/company/tzafon) 3. [Built In – Tzafon Careers](https://builtin.com/company/tzafon) 4. [Tech.eu – Tzafon raises $9.7M pre-seed](https://tech.eu/2025/07/30/tzafon-raises-9-7m-pre-seed-to-scale-ai-compute-and-launch-lightcone/) ## Other roles at Tzafon - [Member of Technical Staff – Backend/Platform Engineer](https://feeny.ai/job/member-of-technical-staff-backend-platform-engineer-tzafon-san-francisco-vn9anxby0wfz) — San Francisco, CA / Tel Aviv, Israel / Zurich, Switzerland - [Member of Technical Staff – Applied AI](https://feeny.ai/job/member-of-technical-staff-applied-ai-tzafon-san-francisco-nmpctyt6w7pv) — San Francisco, CA