Job Title: AI Researcher (AI-Oriented Knowledge Systems)
Location: Piscataway, NJ
The estimated salary range for this role is $85,000 - $145,000 depending on experience.
Responsibilities
Core Research Directions: Responsible for one or two of the following areas:
Knowledge Extraction & Structuring
- Research techniques for extracting structured knowledge from multi-source heterogeneous data (documents, web pages, databases, conversation logs)
- Design automated pipelines for entity recognition, relation extraction, and event detection
- Develop knowledge quality assessment and cleaning mechanisms to filter noise and conflicting information
- Explore LLM-assisted knowledge extraction methods, balancing automation efficiency with manual validation costs
- Research incremental knowledge extraction strategies to support continuous knowledge base updates and expansion
Knowledge Organization & Representation
- Design knowledge graph schemas and ontologies to build structured frameworks for domain knowledge
- Research Knowledge Embedding techniques to achieve fusion of knowledge and vector spaces
- Develop multi-level knowledge representation systems supporting coarse-to-fine granularity knowledge navigation
- Explore knowledge fusion and alignment techniques to resolve entity disambiguation and conflict resolution from multi-source knowledge
- Research knowledge version management and provenance mechanisms to ensure knowledge traceability
Knowledge Retrieval & Augmentation
- Optimize RAG (Retrieval-Augmented Generation) systems to improve retrieval accuracy and answer quality
- Research hybrid retrieval strategies combining vector search, keyword search, graph traversal, and other approaches
- Develop retrieval re-ranking algorithms to enhance Top-K result relevance
- Design retrieval-generation collaborative optimization mechanisms to reduce hallucinations and erroneous citations
- Explore retrieval feedback learning to continuously optimize retrieval strategies based on user behavior
Knowledge Reasoning & Question Answering
- Research knowledge graph-based reasoning techniques supporting multi-hop reasoning, logical reasoning, and causal reasoning
- Develop Complex QA systems supporting multi-condition and multi-step question answering
- Explore fusion methods combining LLMs with symbolic reasoning, leveraging advantages of both neural and symbolic approaches
- Design interpretability frameworks for reasoning processes, supporting answer provenance and reasoning chain visualization
- Research knowledge gap detection and active learning mechanisms to identify coverage blind spots in the knowledge base
Knowledge Update & Maintenance
- Design knowledge timeliness management mechanisms supporting knowledge expiration detection and automatic updates
- Research knowledge conflict detection and resolution strategies for handling contradictory information fusion
- Develop knowledge base health monitoring systems tracking coverage, accuracy, freshness, and other metrics
- Explore human-feedback-driven knowledge iteration mechanisms
- Research knowledge compression and summarization techniques to optimize storage efficiency and retrieval performance
Job Requirements
Basic Qualifications
- Master's degree or above in Computer Science, Artificial Intelligence, Information Management, or related fields
- 3+ years of AI-related research or development experience with hands-on experience in knowledge graphs, RAG, or QA systems
- Publications in top-tier conferences (ACL, EMNLP, SIGIR, WWW, NeurIPS, etc.) preferred
Technical Skills
Programming & Engineering
- Proficient in Python with expertise in data processing and large-scale text processing techniques
- Familiar with mainstream NLP frameworks (spaCy, NLTK, HuggingFace Transformers, etc.)
- Experience with graph databases (Neo4j, NebulaGraph, JanusGraph, etc.)
- Familiar with vector databases (Milvus, Chroma, Weaviate, FAISS, etc.)
AI Expertise
- Deep understanding of core NLP technologies: entity recognition, relation extraction, text classification, semantic similarity
- Familiar with the full lifecycle of knowledge graph construction and application
- Proficient in RAG technology stack with hands-on experience in retrieval optimization, re-ranking, and answer generation
- Prior experience in vertical domain knowledge system construction and knowledge-driven LLM application deployment (e.g., healthcare, legal, finance, technology) preferred
- Familiar with multi-modal knowledge processing (text + image + table + structured data)
Data Processing
- Familiar with large-scale data processing technologies (Spark, Flink, Dask, etc.)
- Experience in data governance such as data cleaning, deduplication, and standardization preferred
- Familiar with common data formats and protocols (JSON, XML, RDF, OWL, etc.)
Research Capabilities
- Ability to conduct independent technical research, owning the full process from problem definition to solution deployment
- Strong literature review and summarization skills with ability to quickly absorb cutting-edge research findings
- Experimental design and evaluation capabilities, able to design proper comparative and ablation studies
Soft Skills
- Strong interest in the intersection of knowledge engineering and AI, keeping up with latest domain developments
- Excellent communication and collaboration skills, able to work efficiently with engineering and product teams
- Systems thinking ability to approach knowledge system design from an overall architecture perspective
#LI-EB1
#GS