3 months ago
Singapore, SingaporeMid Level
Responsibilities
- Design and develop enterprise LLM applications such as intelligent Q&A systems, knowledge-base assistants, AI copilots, document analysis tools, and customer service agents.
- Build end-to-end RAG systems covering document parsing, chunking, embeddings, hybrid retrieval, reranking, and source-attributed response synthesis.
- Develop prompt engineering, context management, structured output, prompt versioning, and reliability strategies for LLM applications.
- Design AI agents using ReAct, Plan-and-Execute, Reflection, function calling, tool use, and multi-agent collaboration patterns.
- Deploy and optimize open-source LLMs in on-premise and private cloud environments, including model serving, quantization, batching, speculative decoding, and tensor parallelism.
- Build model-serving infrastructure, GPU scheduling, AI gateways, fine-tuning pipelines, domain datasets, evaluation benchmarks, and LLM quality-monitoring systems.
- Implement production monitoring, A/B testing, cost and performance tracking, drift detection, security guardrails, PII detection, prompt-injection defense, and content moderation.
- Track frontier AI research and contribute to internal technical documentation, talks, and best-practice guides.
Requirements
- Bachelor's degree or above in Computer Science, Artificial Intelligence, Machine Learning, or a related technical field; a master's degree or PhD is preferred.
- At least 2 years of professional AI/ML engineering experience, including production deployment of LLM systems at scale.
- Deep understanding of Transformer architecture, attention mechanisms, and LLM pre-training, fine-tuning, and inference.
- Expertise with production LLM application frameworks such as LangChain, LlamaIndex, or Haystack.
- Hands-on experience with RAG, vector databases, embedding models, reranking, hybrid search, query expansion, and HyDE.
- Experience deploying and tuning DeepSeek, Qwen, Kimi, Llama, Mistral, Mixtral, Gemma, or Phi models.
- Strong experience with model-serving infrastructure, GPU scheduling, quantization, inference optimization, KV-cache optimization, and memory-efficient attention.
- Strong Python programming skills and experience with PyTorch, TensorFlow, or JAX; familiarity with FastAPI or Flask for LLM API services.
- Experience with LLM evaluation, A/B testing, and production monitoring of AI systems.
- Preferred experience includes LangGraph, AutoGen, CrewAI, OpenAI Assistants API, MCP, multimodal LLMs, DSPy, PromptLayer, LangSmith, model distillation, cloud GPU providers, open-source AI contributions, or NLP/LLM publications.
Tech Stack
Categories
About AvePoint
AvePoint builds SaaS backup, migration, and governance software for enterprises running Microsoft 365, SharePoint, Google Workspace, and Salesforce. Its products help secure data, manage compliance, and automate administration, sold via subscriptions and channel partners. Founded in 2001 and headquartered in Jersey City, the company is publicly traded on the NASDAQ.