LLM Algorithm Applications Engineer

【About Us】

• VICI Holdings is a top-tier company focused on cutting-edge technology and financial trading. Our hardware team is dedicated to developing ultra-low latency and high-performance digital design solutions, keeping pace with Wall Street’s best in FPGA design. Our strategies span stocks, futures, and derivatives, with a daily global trading volume reaching hundreds of millions of USD, underscoring our leadership and scale.
• Engineering-driven trading: we turn research into quantifiable trading edge through robust data and systems engineering.
• Global perspective with local agility: Taipei-based team operating across global markets with high-speed decisioning and strong compliance.
• （HPC/GPU/FPGA）。 Leading infrastructure: proprietary low-latency trading stack, distributed data pipelines, and HPC/GPU/FPGA acceleration.
• Culture & values: Ownership, performance obsession, transparent collaboration, fast iteration, and outcomes-first mindset.
• Growth & resources: generous experiment budgets, top-tier cloud/compute, and support for academic and open-source engagement.

【About The Position and Team】

• This role builds and strengthens our core capabilities in Large Language Model (LLM) Agents and Reinforcement Learning (RL) to directly power trading automation and deepen market insights.
• You will work closely with quants, data engineers, and software engineers to turn research prototypes into production-grade systems for live trading environments.

【R&R】

• Co-design experiments with quants to translate model outputs into tradable signals, with offline/online backtesting and A/B testing.
• Design and build AI agents (RAG, tool use, task planning) to boost strategy research and trading execution efficiency.
• Apply prompt engineering, MoE, LoRA, quantization, and distillation to improve the perf–cost curve and inference latency.
• Specialize fine-tuning and alignment (SFT/DPO) for financial text—disclosures, news, social, earnings—to improve event and sentiment inference.
• Adopt MLOps best practices (containerization, CI/CD, feature/model versioning) to increase velocity and compliance.
• Track frontier research and lead internal tech sharing to inform model selection and architecture evolution.

【 Requirements】

• M.S. or above in CS/EE/Data Science preferred; equivalent practical achievements considered.
• Proficient with AI coding tools (Cursor, Claude Code); adept in spec engineering and test-driven development (TDD). Deliver rapid 0→1 with AI and drive 1→100 iterative improvement.
• 2+ years in LLM with PyTorch or TensorFlow; hands-on experience training and fine-tuning models end to end.
• Fluency in the LLM ecosystem (LangGraph, Ollama, OpenAI SDK) to rapidly prototype and ship services.
• Strong experiment design, problem decomposition, and cross-team communication with an emphasis on observability and operability.
• Effective communication in Mandarin and English; able to write technical docs and present externally.

【Other Requirements】

• Financial experience (market prediction, event extraction, sentiment analysis, factor modeling, portfolio optimization).
• Competitions/community (Kaggle, AI Challenger) or active contributor to open-source projects.
• Publications or talks at top venues (NeurIPS/ICML/ICLR/CVPR).

【Character】

• Hands‑on: builds and ships; outcome‑oriented.
• Goal‑driven & resilient: sets a high bar and delivers under pressure and ambiguity.
• Passion for AI × markets: sustained curiosity and commitment to application.
• Bias to action: experiment, measure, iterate quickly.
• Owner mindset: accountable end‑to‑end; proactively aligns stakeholders.
• Low‑ego, direct communication: candid feedback and fact‑based discourse.