Senior AI Engineer - Agentic Data

Beyond is Qodea’s Customer Experience Design Studio.

We design the ‘surfaces’ where customers and technology meet.

Our teams shape the intelligence behind those experiences, turning data, design, and emerging technologies into products that are intuitive, adaptive, and human.

We are multi-disciplinary designers, product strategists, writers, architects, engineers, data scientists, and ML researchers, united by a single goal: to design a better future for our clients and their customers.

We believe we are on the cusp of a new golden era of design, one where design will be more important than ever. An era of exploration and discovery.

We’re building a studio where designers immerse themselves in AI design paradigms, experimenting with adaptive patterns, conversational interfaces, and agentic workflows, the foundation for tomorrow’s customer experience.

We look for people who embody:

Innovation to solve the hardest problems.

‍Accountability for every result.

‍Integrity always.

About The Role:

• We’re seeking a versatile and expert Senior AI Engineer to join our specialized Agentic Pods as the primary owner of data intelligence. In this role, affectionately dubbed "The Librarian," you will ensure that our autonomous agents are as intelligent as the data they access. As we shift from explicit rules to orchestrating cognitive architectures, your mission is to make existing enterprise data "agent-ready."

• Working alongside the Lead AI Engineer within a high-traffic, commercial-grade environment, you will define the retrieval strategies that power secure and performant agentic experiences. You will move beyond simple data storage to building dynamic context engines, ensuring agents can retrieve the correct pricing, policy, or product information with zero hallucination. You will bridge the gap between traditional search (Elastic) and cognitive retrieval (Vector/Graph), designing the "long-term memory" for our digital solutions.

What You'll Do:

• Architect Advanced Retrieval Strategies: Design and implement robust RAG (Retrieval Augmented Generation) pipelines, deciding on optimal "chunking" strategies for complex unstructured data (websites, policy documents) to maximize LLM comprehension.
• Drive Graph vs. Vector Decisions: Lead the comparative analysis and implementation of standard Vector RAG versus GraphRAG. You will determine when to use semantic similarity versus knowledge graph traversals to ground agent reasoning in factual reality.
• Build "Agent-Ready" Data Pipelines: Create ingestion pipelines using Python and Google Cloud technologies that transform raw data a diverse set of sources including websites, online drives (e.g. Google Drive), databases and other platforms (e.g. ElasticSearch) into high-quality vector embeddings and knowledge graph entities.
• Ensure Data Hygiene & Relevance: Implement mechanisms to ensure agents always retrieve the most current state of the world (e.g., current stock vs. last year's cache), enforcing strict data contracts to prevent stale context injection.
• Manage Vector & Graph Stores: Deploy and optimize tools like Vertex AI Vector Search and graph databases, managing the indexing, retrieval latency, and cost-performance trade-offs of the agent's memory.
• Orchestrate Context Building: Develop the logic that dynamically assembles the "context window" for the Lead AI Engineer's agent orchestration, balancing token limits with information density.
• Implement Observability for Retrieval: detailed logging of retrieval scores and re-ranking logic to allow for "Trace Reviews," ensuring we can debug why an agent selected a specific piece of information.

Requirements:

• 6+ years of engineering experience, with a focus on Data Engineering, Machine Learning, or Backend systems.
• Applied GenAI Experience: At least 2 years of hands-on experience building RAG architectures, specifically focusing on the data ingestion and retrieval side (GCP ADK, LlamaIndex, LangChain).
• Vector Database Mastery: Deep knowledge of Vertex AI Vector Search, Pinecone, or Weaviate, including expertise in HNSW indexing, dimensionality reduction, and hybrid search (keyword + semantic).
• GCP Ecosystem fluency: Strong experience with Google Cloud Platform, specifically BigQuery, Dataflow (Apache Beam), and Cloud Functions for serverless data processing.
• Practical Graph Experience: Proven ability to implement Property Graphs using Neo4j (Cypher) or NetworkX. You understand how to extract entities and relationships from unstructured text to build a "Knowledge Graph" without getting bogged down in academic ontologies.
• Search Expertise: Experience working with ElasticSearch or OpenSearch, particularly in hybrid configurations where vector search is augmented by lexical filtering.
• Python Proficiency: Mastery of Python for data manipulation (Pandas, Polars) and the core AI ecosystem (SentenceTransformers, OpenAI API, PyTorch).
• Data Quality Obsession: A track record of implementing automated data validation and "Golden Datasets" to grade retrieval accuracy before it reaches the LLM.

Preferred Qualifications:

• Experience with "Eval-Driven Development" using frameworks from Google ADK, Ragas, TruLens, or DeepEval to quantitatively measure RAG performance (Context Precision, Context Recall).
• Familiarity with Agentic Orchestration frameworks like Google ADK or LangGraph, and understanding how agents utilize "tools" to query databases.
• Experience building MCP (Model Context Protocol) servers or similar gateways to expose data to agents.
• Background in semantic web technologies (RDF, SPARQL) applied in a pragmatic, modern context.

Diversity and Inclusion

At Beyond, we champion diversity and inclusion. We believe that a career in IT should be open to everyone, regardless of race, ethnicity, gender, age, sexual orientation, disability, or neurotype. We value the unique talents and perspectives that each individual brings to our team, and we strive to create a fair and accessible hiring process for all.