ML Research (junior-senior)

ML Research (junior-senior)

Higgsfield AI, 100-350k$/year, office in Almaty with subsequent relocation to the USA, full relocation support

Higgsfield AI is one of the fastest-growing GenAI companies in the world. We are currently building a new research laboratory completely focused on Diffusion Models, Multimodal AI, and Reinforcement Learning. We are starting with a budget of $100M and scaling up to $1B.

Currently, entering the LLM frontier costs hundreds of billions of dollars, whereas in diffusion models, the industry is just beginning, and a foundation laboratory can be built from scratch. Neither OpenAI (Sora), nor Google (Veo), nor Meta has managed to surpass Chinese companies. At the moment, the frontier is open, and we see an opportunity to create the best research laboratory in the world.

We believe that the next leap in video generation will come through Reinforcement Learning on top of diffusion transformers, just as RLVR changed the development of LLMs.

We are creating the first such laboratory in the Russian-speaking space. Today, leading research centers are located in the USA and China, while there is a strong community of Russian-speaking researchers that is not yet united around its own frontier laboratory.

Responsibilities

• Research architectures for Diffusion Models, Multimodal AI, and MoE.
• Work with Reinforcement Learning for generative models.
• Design fault-tolerant training pipelines with maximum GPU utilization on large clusters.
• Train models for filtering and evaluating multimodal data for training.
• Accelerate inference and improve models using distillation, alignment, and quantization.

We are looking for

• Experience in Diffusion Models, Multimodal AI, or Reinforcement Learning.
• Strong command of PyTorch, including torch.distributed.
• Experience with distributed training, large models, or modern ML stack (DeepSpeed, FSDP, etc.).
• Passion for research and a desire to turn ideas into working models.

Will be a big plus

• Publications at NeurIPS, ICML, ICLR.
• Education or graduation from HSE Lyceum, Faculty of Computer Science HSE (PMI) or other strong technical schools.
• Participation in olympiads or a strong research and engineering background.

Conditions:

• office in Kazakhstan, Almaty
• subsequent relocation to the USA
• junior-senior salary range 100-350k USD/year

It is not necessary to meet every point. If you have a strong mathematical foundation, research mindset, and a desire to build the next generation of ML models — we would be glad to meet you.