5 hours ago
Taipei, Taiwan +2 moreIntern
Responsibilities
- Design and develop LLM-powered recommendation and personalization systems, including candidate generation, ranking, reranking, intent understanding, and context-aware recommendation.
- Build agentic AI systems using internal data, APIs, tools, and domain-specific capabilities for complex financial and trading tasks.
- Develop tool routing, tool retrieval, planning, and multi-step reasoning mechanisms for large-scale tool ecosystems.
- Perform LLM post-training, including supervised fine-tuning, preference optimization, and reinforcement learning.
- Build training and evaluation datasets, benchmarks, and pipelines measuring relevance, personalization, tool selection, task success, latency, and reliability.
- Prototype and integrate LLM and agent workflows covering retrieval, recommendation, planning, execution, verification, memory, and feedback loops.
- Explore small and specialized language models for routing, recommendation, classification, reranking, and latency-sensitive tasks.
- Collaborate with engineers, researchers, product teams, and domain experts on system design, experimentation, integration, deployment, and optimization.
Requirements
- Strong foundation in machine learning, NLP, information retrieval, recommendation systems, or large language models.
- Hands-on experience with LLM post-training, recommendation and ranking systems, LLM agents and tool-use systems, Retrieval-Augmented Generation, or reinforcement learning and preference optimization.
- Familiarity with SFT, RL, DPO/GRPO-style optimization, prompt engineering, structured generation, function or tool calling, and model evaluation.
- Good understanding of embedding-based retrieval, learning-to-rank, reranking, personalization, user modeling, or generative recommendation.
- Strong Python programming skills and experience with deep learning frameworks such as PyTorch.
- Ability to independently conduct experiments, analyze model or system performance, and translate research ideas into production solutions.
- Preferred: experience with production-scale recommendation, search, or LLM systems; agent frameworks; tool routing; multi-agent or memory systems; long-horizon agent workflows; LLM inference and serving; large-scale datasets; distributed training; model serving; high-throughput online systems; Web3, cryptocurrency, financial markets, or trading systems.
- Publications, open-source contributions, or practical projects related to LLMs, recommendation systems, agents, search, or reinforcement learning are a plus.
Benefits
- 3-6 month internship program with networking, professional development, career growth, and continuous learning opportunities.
- Work-from-home arrangement, which may vary based on the business team's work.
- Competitive salary and company benefits.
- Engagement is subject to contract and applicable local laws.
About Binance
Binance is the world’s leading blockchain ecosystem and cryptocurrency infrastructure provider with a product suite that includes the world's largest digital asset exchange and much more. Trusted by over 200 millions of users worldwide, the Binance platform is dedicated to increasing the freedom of money for users, and features an unmatched portfolio of crypto products and offerings, including: trading and finance, education, data and research, social good, investment and incubation, decentralization and infrastructure solutions, and more. Posts are not directed towards UK users.