UST
Website:
ussmokeless.com
Company:
https://www.linkedin.com/company/ustglobal
Seniority: Not Applicable
Industries: IT Services and IT Consulting
Job details:
Role Description
Job Title: AI Engineer
Role Summary:
The AI Engineer will design, build, and own the AI-powered applications and data enrichment services that put Client’s product data to work, including retrieval-augmented generation (RAG), lexical and semantic search, and agent-based solutions.
This role is hands-on and delivery-oriented, working primarily in Google Cloud Platform (GCP) with Vertex AI, AlloyDB, BigQuery, and Cloud Run as core components of the stack.
The AI Engineer will apply LLMs during ingestion to improve product data quality and will build the agents and search experiences that use that data.
We are looking for an engineer who has put AI solutions into production, measured the quality of model output rather than assuming it, and can speak candidly about what broke, how it was diagnosed, and what changed as a result.
Reporting to the Manager of the AI Technology and Innovation Platform team, this position will collaborate closely with the AI Architect, Data Engineers, Platform Engineer, QA, and business stakeholders to move solutions from experimentation to secure, reliable, and measurably valuable production systems.
Duties And Responsibilities:
Design, build, and deploy AI agents using frameworks such as Google Agent Development Kit (ADK), LangGraph, or LangChain, including tool/function calling, state and memory, orchestration, failure handling, and guardrails, following patterns defined by the AI Architect
Design, develop, and deploy RAG pipelines, including document and product data ingestion, chunking, embedding generation, retrieval, re-ranking, and grounded response generation
Build lexical, semantic, and hybrid search solutions that combine full-text/BM25-style ranking with vector similarity, and tune them for relevance
Build LLM-based data enhancement during ingestion (e.g., attribute extraction, categorization, normalization, deduplication, and enrichment of product records), with schema validation, confidence thresholds, and review queues for low-confidence results
Develop and tune data and retrieval layers in AlloyDB (schema design, pgvector/ScaNN indexing, full-text search, query tuning) and in BigQuery (analytics, batch enrichment, and vector search where appropriate)
Build and deploy services on Cloud Run (services and jobs) and integrate with Pub/Sub, Dataflow, and Cloud Composer for event-driven and batch ingestion, with idempotency, retries, dead-letter handling, and safe reprocessing
Build, integrate, and operate solutions on Vertex AI, including Gemini and other foundation models, embeddings, Vertex AI Search, and Agent Builder/Agent Engine
Create and run evaluation processes for retrieval, enrichment, and agent quality (golden datasets, sampling, precision/recall, groundedness, relevance, cost per record, latency), including automated regression checks when prompts or models change, in partnership with QA
Implement controls that prevent bad data from reaching the product catalog, such as approval gates, dry-run modes, audit trails, and constrained agent permissions
Take end-to-end ownership of what you ship, including monitoring, ing, root cause analysis, and post-incident reviews
Build and expose REST/gRPC APIs and integrate with enterprise systems and Google interfaces used by AI applications
Work with the Platform Engineer on CI/CD, containerization, and Infrastructure as Code for deployment
Apply security, privacy, and responsible AI practices, including IAM, secrets management, data handling, and auditability
Monitor and optimize performance and cost, including token usage, query cost, and compute
Participate in design and code reviews, and document designs, runbooks, and evaluation results
Qualifications And Requirements:
Bachelor’s degree in Computer Science, Engineering, Data Science, or a related field, or equivalent practical experience
4+ years of software or data engineering experience, with 2+ years building LLM-based applications
Strong Python skills, including API development (FastAPI, Flask, or similar) and writing testable, maintainable code
Hands-on experience with LLM APIs (Vertex AI/Gemini, OpenAI, Anthropic, or similar) in real workloads, including prompt design and output validation
Practical experience evaluating LLM output, with the ability to describe the methods, metrics, and tradeoffs used
Hands-on experience building RAG solutions, including embeddings, vector search, chunking, and re-ranking
Experience implementing semantic and lexical search, including hybrid retrieval
Experience on Google Cloud Platform, including Vertex AI, BigQuery, and containerized services such as Cloud Run or GKE
Advanced SQL and experience with AlloyDB or PostgreSQL, including pgvector
Demonstrated experience with schema and data pipeline design, including idempotency, backfills, and handling schema changes
Experience building at least one agent or multi-step LLM workflow that ran in production or in a real pilot, with an explanation of its design, testing, and failure handling
Proven ability to discuss production incidents: what went wrong, how it was diagnosed and fixed, and what was changed to prevent recurrence
Familiarity with Git, CI/CD, Docker, and cloud security and IAM fundamentals
Strong communication skills and the ability to work across technical and business teams
Skills
AlloyDB, LangChain, LangGraph, Google Cloud Platform, RAG, Vertex AI, BigQuery, Generative AI, Agentic AI, Python, Prompt Engineering
Skills
AlloyDB, LangChain, LangGraph, Google Cloud Platform, RAG, Vertex AI, BigQuery, Generative AI, Agentic AI, Python, Prompt Engineering
Click on Apply to know more.