Workato
About this role
One of the more technically interesting roles in today's list and, at three to five years, one of the most reachable. You would build the retrieval layer behind Workato's enterprise AI agents: the system that answers a question by stitching together a Salesforce record, a call summary, a Zendesk ticket, a Jira blocker and a SharePoint contract into one ranked, permission aware response. The posting is honest about why that is hard. Every source has its own access control model, the content types are wildly different, and freshness matters because yesterday's answer is often wrong. The work combines classical information retrieval, dense vector search, knowledge graphs and LLM reasoning. It is explicitly described as a hands on senior individual contributor role. Hyderabad.
Who this is for
What the posting asks for:
- 3 to 5 years building production search, retrieval, knowledge base or recommendation systems.
Day to day:
- Building a unified retrieval layer across enterprise systems including Google Drive, SharePoint, Confluence, Jira, Asana, Zendesk, Freshdesk, Salesforce and Notion, exposed through a clean agent friendly interface.
- Designing hybrid retrieval pipelines that combine lexical search (BM25), dense vector retrieval and structured SQL or graph retrieval, with re-ranking tuned for cross source results.
- Engineering ingestion and freshness pipelines that incrementally sync millions of documents, tickets, tasks and CRM records at low latency and predictable cost.
- Shipping production grade observable systems with strong service level objectives on latency, freshness, recall and cost, plus the dashboards and tracing to evidence them.
The genuinely hard parts, named by the employer: heterogeneous content types, heterogeneous permission models where every source has its own access control list, and real freshness constraints.
Scope: described in the posting as a hands on senior individual contributor role for someone who wants to go deep on retrieval quality.
Location and office reality: Hyderabad. No office day policy is published on this posting, and unlike Workato's Customer Centric Engineer role there is no night shift stated here.
Honest fit guidance: if you have built search or RAG systems that ran in production and you can talk about recall and ranking quality with numbers, three years is enough here. If your retrieval experience is a demo or a tutorial pipeline, the bar in the interview will be evidence of production quality work.
- 3 to 5 years building production search, retrieval, knowledge base or recommendation systems.
Day to day:
- Building a unified retrieval layer across enterprise systems including Google Drive, SharePoint, Confluence, Jira, Asana, Zendesk, Freshdesk, Salesforce and Notion, exposed through a clean agent friendly interface.
- Designing hybrid retrieval pipelines that combine lexical search (BM25), dense vector retrieval and structured SQL or graph retrieval, with re-ranking tuned for cross source results.
- Engineering ingestion and freshness pipelines that incrementally sync millions of documents, tickets, tasks and CRM records at low latency and predictable cost.
- Shipping production grade observable systems with strong service level objectives on latency, freshness, recall and cost, plus the dashboards and tracing to evidence them.
The genuinely hard parts, named by the employer: heterogeneous content types, heterogeneous permission models where every source has its own access control list, and real freshness constraints.
Scope: described in the posting as a hands on senior individual contributor role for someone who wants to go deep on retrieval quality.
Location and office reality: Hyderabad. No office day policy is published on this posting, and unlike Workato's Customer Centric Engineer role there is no night shift stated here.
Honest fit guidance: if you have built search or RAG systems that ran in production and you can talk about recall and ranking quality with numbers, three years is enough here. If your retrieval experience is a demo or a tutorial pipeline, the bar in the interview will be evidence of production quality work.
Apply on company site
Opens www.workato.com, the employer's own application page. Applying is always free.