Visor.ai

• Design and develop end-to-end NLP pipelines — from classical text processing to state-of-the-art LLM-powered architectures • Build and maintain systems for intent detection, NER, entity extraction, and text classification, both standalone and as components feeding into larger LLM workflows • Design and optimize Retrieval-Augmented Generation (RAG) systems — chunking strategies, vector store architecture, hybrid search (dense + sparse), and re-ranking pipelines • Work with embedding models for semantic search, document retrieval, and intent classification in contact center contexts • Design and implement agentic architectures — tool use, function calling, multi-step reasoning, and orchestration with frameworks like LangChain, LlamaIndex, or custom-built solutions • Develop memory and context management strategies — short-term conversation memory, long-term user context, and context window optimization for multi-turn interactions • Evaluate and benchmark models rigorously: hallucination detection, faithfulness scoring, latency/token cost tradeoffs, and continuous performance monitoring • Integrate AI components into scalable, production-ready microservices with a focus on low-latency inference pipelines • Collaborate with product and engineering to design new AI-powered features and drive innovation across the platform

• Design and develop end-to-end NLP pipelines — from classical text processing to state-of-the-art LLM-powered architectures • Build and maintain systems for intent detection, NER, entity extraction, and text classification, both standalone and as components feeding into larger LLM workflows • Design and optimize Retrieval-Augmented Generation (RAG) systems — chunking strategies, vector store architecture, hybrid search (dense + sparse), and re-ranking pipelines • Work with embedding models for semantic search, document retrieval, and intent classification in contact center contexts • Design and implement agentic architectures — tool use, function calling, multi-step reasoning, and orchestration with frameworks like LangChain, LlamaIndex, or custom-built solutions • Develop memory and context management strategies — short-term conversation memory, long-term user context, and context window optimization for multi-turn interactions • Evaluate and benchmark models rigorously: hallucination detection, faithfulness scoring, latency/token cost tradeoffs, and continuous performance monitoring • Integrate AI components into scalable, production-ready microservices with a focus on low-latency inference pipelines • Collaborate with product and engineering to design new AI-powered features and drive innovation across the platform • Conduct practical research with a scientific mindset and a focus on delivery — publications are possible and encouraged

• Designing, planning, implementing, testing, and analyzing production code • Monitoring and providing support for systems in production • Implementing creative solutions to complex problems, using AI-assisted approaches where appropriate • Developing and maintaining data infrastructure and computer systems • Ensuring development is performance-, scalability-, and security-focused • Creating and managing NoSQL databases • Managing and maintaining servers