BlueCloud
Website:
blue.cloud
Job details:
Overview
We are seeking a specialized AI/ML Engineer/ Architect to spearhead the development of a conversational AI interface using Snowflake Cortex. This role focuses on the "Intelligence" layer of our POC—implementing advanced Retrieval-Augmented Generation (RAG) and agentic workflows to transform raw business data into high-fidelity insights for PMs, Sales, and Marketing teams.
Key Responsibilities:
• AI Architecture & Strategy: Lead the design of the Cortex Search strategy, selecting optimal embedding models and chunking strategies to maximize retrieval precision.
• Advanced RAG Implementation: Build and optimize RAG workflows using Snowflake’s native AI functions; refine the "Retrieval" stage through hybrid search (semantic + keyword) and re-ranking techniques.
• Agentic Framework Development: Design and implement Cortex Agents using the Planner-Tool-Executor-Refiner framework to handle complex, multi-step GTM queries.
• Prompt Engineering & Model Tuning: Develop and version-control system prompts; utilize Snowflake Cortex functions for text summarization, translation, and sentiment analysis to enrich the user experience.
• LLM Evaluation (LLM-as-a-Judge): Establish an evaluation framework to measure the quality of AI responses (faithfulness, relevancy, and toxicity) and monitor for hallucination.
• Compute & Inference Optimization: Manage and optimize the use of Snowflake’s AI-ready compute, ensuring a balance between model performance (latency) and credit consumption.
Required Skills & Experience
• 8+ years in AI, ML, or Data Engineering, with at least 2 years focused on LLM application development.
• Snowflake Cortex Mastery: Deep hands-on experience with CORTEX.COMPLETE, CORTEX.EMBED_TEXT, and Cortex Search.
• Vector Mechanics: Strong understanding of vector databases, cosine similarity, and high-dimensional indexing within Snowflake.
• RAG Expertise: Proven experience in chunking strategies (recursive, fixed-size) and managing the "context window" for LLMs.
• Python & SQL Proficiency: Ability to write sophisticated Python UDFs and Snowpark code for ML preprocessing.
• Unstructured Data Science: Experience transforming raw, noisy text (Gong transcripts, PDFs) into clean, machine-readable formats for AI consumption.
• Agentic Knowledge: Understanding of how to build "Loop-based" AI systems that can use tools (SQL API, web search, etc.) to solve problems.
Click on Apply to know more.