LLM Data Strategy Expert (Annealing / SFT)

Bitdeer AI Lab Position

Bitdeer is a world-leading technology company for Bitcoin mining and AI cloud. Bitdeer is committed to providing comprehensive Bitcoin mining solutions for its customers. Apart from designing industry-leading ASIC chips and manufacturing mining rigs, the Group handles complex processes involved in computing across the value chain. This includes equipment procurement, transport logistics, datacenter design and construction, equipment management, and network and facility operations. Bitdeer also offers advanced cloud capabilities to customers with a high demand for artificial intelligence. Headquartered in Singapore, Bitdeer operates globally with a diversified 3 GW energy portfolio, and deploys Bitcoin mining and HPC datacenters in the United States, Bhutan, Norway, Canada, Malaysia, and Ethiopia.

Bitdeer AI Lab is a frontier AI lab under Bitdeer, a global-leading computing power solutions provider. Guided by long-termism, we are committed to exploring the frontiers of artificial intelligence with the ambition, courage, and determination to build technologies that can truly change the world.

Our mission is twofold: first, to effectively transform energy into intelligence; second, to push the limits of intelligence by rethinking AI systems and architectures that can learn more efficiently, reason more deeply, and scale more effectively.

Our vision is to create intelligence that learns more like humans do: efficiently, adaptively, and recursively, turning finite parameters and finite compute into unbounded potential. We pursue this work with a deep sense of purpose, believing that the most meaningful advances in AI will not only push the frontier of research, but also reshape the future of the world.

Our lab is equipped with thousands of cutting-edge GPUs dedicated to AI research, and we are committed to continuously investing in and expanding our computational infrastructure to support world-class research and engineering in artificial intelligence.

The core mission of this role is to orchestrate the fine-grained reshaping of model capabilities through the design of Task Mixtures and Curriculum Learning, and to establish causal attribution between data decisions and training dynamics.

Responsibilities:

• SFT Architecture Design: Define Task Mixtures and difficulty gradients (Curriculum) across reasoning, coding, and Agent domains. Proactively account for the impact of SFT on the signal-to-noise ratio in RLHF/DPO to prevent capability overfitting.
• Annealing Strategy: Design the data mixture for the annealing phase, optimizing the proportion of high-density task data. Balance generalizability with specialized capabilities and quantitatively monitor Capability Drift.
• Synthetic Data Amplification: Build synthetic data pipelines for verifiable domains (e.g., Math/Code). Design execution-based verifiers to drive the scaling of model capabilities.
• Closed-Loop Attribution System: Establish a robust feedback loop: Data Distribution → Training Dynamics → Model Behavior. Accurately pinpoint data noise or distribution collapse from Eval Regressions.

Hands-on Expertise: Proven experience in designing SFT Mixtures, with a deep understanding of how data distributions impact surface-level model behaviors.

Data Intuition: Sharp intuition for identifying Reasoning Hallucinations and Distribution Collapse; ability to reverse-engineer data mixture issues from loss spikes/fluctuations.

Engineering Foundation: Proficiency in large-scale data processing and inference pipelines (e.g., vLLM, SGLang, Ray).

Bonus Points: Proven track record in implementing Annealing strategies, practical insights into DPO data evolution, or in-depth research in Data-centric AI.

A culture that values authenticity and diversity of thoughts and backgrounds;

An inclusive and respectable environment with open workspaces and exciting start-up spirit;

Fast-growing company with the chance to network with industrial pioneers and enthusiasts;

Ability to contribute directly and make an impact on the future of the digital asset industry;

Involvement in new projects, developing processes/systems;

Personal accountability, autonomy, fast growth, and learning opportunities;

Attractive welfare benefits and developmental opportunities such as training and mentoring.