TL;DR
AI Algorithm Engineer (Web3): Architecting AI Agent systems and optimizing end-to-end RAG pipelines leveraging the largest proprietary dataset in the Web3 field with an accent on ingestion, chunking, embedding, and hybrid vector search. Focus on implementing LLM training/alignment, and deploying scalable for intelligent search and task execution agents.
Location: Global / Hong Kong / Singapore / Dubai
Company
hirify.global is building the most advanced AI Agent for the Web3 industry, leveraging the largest proprietary dataset in the field.
What you will do
- Develop AI Agent Systems: Build intelligent search and task execution agents using ReAct, planning, and multi-agent frameworks.
- Optimize End-to-End RAG Pipelines: Build and refine efficient RAG systems from ingestion, chunking, and embedding to hybrid vector search, implementing precise grounding and citation.
- LLM Training & Alignment: Conduct advanced post-training and align models for reliable JSON-schema function calling and external tool usage.
- Automated Evaluation & Iteration: Build offline/online evaluation pipelines using synthetic QA, retrieval metrics, and hallucination detection to continuously improve system performance and stability.
Requirements
- Bachelor's or Master's degree in Computer Science, AI, Machine Learning, or a related field
- 3+ years of experience developing AI systems, with a focus on RAG, Agent architectures, or LLM training/optimization
- Proficiency in Python and key ML frameworks (PyTorch/TensorFlow), with experience in distributed training and high-performance inference
- Hands-on, in-depth experience in at least two of the following domains: End-to-end RAG pipeline development and optimization with OpenSearch/vector databases, AI Agent framework development, Advanced LLM training and alignment techniques
- Excellent problem-solving and systems thinking skills.
Nice to have
Будьте осторожны: если вас просят войти в iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого - это мошенники. Обязательно жмите "Пожаловаться" или пишите в поддержку. Подробнее в гайде →