数据/情境工程师
Data / Context Engineer
职位名称：数据/上下文工程师
地点：远程
公司简介
Glint Tech Solutions 是一家由女性拥有的全球IT人才派遣和招聘公司，为美国和加拿大的企业客户提供支持。
项目描述
一家领先的企事业单位客户，是一家跨国电信技术公司，正在寻找一名数据/上下文工程师，以在大规模多区域运营中塑造知识流动方式。这是一个高可见度的机会，负责设计和扩展知识库（KB）系统，该系统将直接推动12个地区及更多地区的决策。该职位位于AI基础设施、数据工程和实际业务影响的交汇点，从第一天起就能在全美范围内产生影响。
主要职责
- 与SA合作，在P1阶段签署并实现多区域KB架构
- 在P2阶段为12个地区建立KB（第一组第1周，第二组第2周，第三组第3周的7月）
- 与采样人类审批流程一起构建和运营摄入管道（机器可读的区域标准嵌入检索模式）
- 在9月至10月期间添加特定于MOD的可检索上下文作为区域覆盖层
- 负责KB完整性检查、检索评估和每周健康报告
必备技能
- 4-6年在生产环境中实际操作RAG / 检索 / 向量存储工程的经验
- 具有构建多租户或跨区域检索架构的经验，具有覆盖/继承语义，而不仅仅是“一个大索引”
- 熟悉Vertex AI Vector Search 或可转移的深度（Pinecone、Weaviate、pgvector 与强大的租户功能、OpenSearch 混合）
- 嵌入模型评估的专业性——检索质量指标（recall@k、precision@k、MRR），而不是感觉
- Python；熟悉结构化文档摄入管道（PDF / XML / 电子表格分块、标准化、嵌入）
加分技能
- 设计了草稿/批准/退役模式之间的晋升路径，并带有审计日志
- 与写回路径集成的采样人类审核工作流
- 熟悉RF / 电信标准格式
最初发布于Himalayas
查看英文原文
Job Title: Data / Context Engineer
Location: Remote
Company Overview
Glint Tech Solutions is a women-owned, global IT staffing and recruiting firm supporting enterprise clients across the USA and Canada.
Project Description
A leading enterprise client, a multinational telecommunications technology company, is seeking a Data / Context Engineer to play a pivotal role in shaping how knowledge flows across a massive, multi-region operation. This is a high-visibility opportunity to architect and scale a knowledge base (KB) system that will directly power decision-making across 12 regions and beyond. The role sits at the intersection of AI infrastructure, data engineering, and real-world business impact, with contributions visible across a nationwide operation from day one.
Key Responsibilities
- Sign and implement the multi-regional KB architecture in P1 alongside SA
- Seed the KB across 12 regions during P2 (cohort 1 wk1, cohort 2 wk2, cohort 3 wk3 of July)
- Build and operate the ingestion pipeline (machine-readable regional standards embeddings retrievable patterns) with sampled human approval gate
- Add MOD-specific retrievable context as regional overlays during Sep-Oct
- Own KB integrity checks, retrieval evaluation, and weekly health reporting
Mandatory Skills
- 4-6 years of hands-on RAG / retrieval / vector store engineering in production
- Experience building a multi-tenant or multi-region retrieval architecture with overlay / inheritance semantics, not just "one big index"
- Vertex AI Vector Search or transferable depth (Pinecone, Weaviate, pgvector with strong tenancy, OpenSearch hybrid)
- Embedding model evaluation discipline — retrieval quality metrics (recall@k, precision@k, MRR), not vibes
- Python; familiar with structured-doc ingestion pipelines (PDF / XML / spreadsheet chunked, normalized, embedded)
Nice-to-Have Skills
- Designed a promotion path between draft/approved/retired patterns with audit log
- Sampled human-review workflows integrated with the writeback path
- RF / telecom standards format familiarity
Originally posted on Himalayas