Website:
dialnexa.com
Job details:
ABOUT THE ROLE
We are building two things that will define how DialNexa operates as a company and this engineer will own both of them.
The first is a company-wide knowledge base that makes everything DialNexa knows, about its product, clients, processes, data, and decisions, queryable by AI agents. Think of it as the memory layer of the company.
The second is the DialNexa AI OS, a growing suite of internal AI agents that automate work across operations, marketing, sales, engineering, and product. Every function in the company will have agents doing work that would otherwise require headcount. This engineer builds and maintains those agents.
This is not a research role. This is not a role where you fine-tune models and write papers. This is a deeply practical, execution-heavy role for someone who has already built AI agents that work in production and wants to build an entire operating system for a company from scratch.
If the idea of being the person who makes a whole company run leaner through AI genuinely excites you, this role was written for you.
WHAT YOU WILL OWN
Company Knowledge Base
- Design and build a structured, queryable knowledge base that ingests data from across the company including product docs, client data, call transcripts, internal processes, CRM data, and more
- Choose and manage the right vector database infrastructure to power fast, accurate retrieval
- Build ingestion pipelines that keep the knowledge base current without manual effort
- Make sure every AI agent in the company can pull the right context at the right time
DialNexa AI OS
- Build and maintain AI agents that automate real work across every function in the company
- Operations agents that monitor client health, flag risks, and surface insights from dashboards
- Marketing agents that research prospects, draft content, and run enrichment workflows
- Sales agents that qualify leads, prepare call briefs, and follow up on pipeline
- Engineering and product agents that summarise tickets, generate drafts, and reduce coordination overhead
- Own the agent architecture end to end, prompting, tool use, memory, orchestration, and reliability
Infrastructure and Reliability
- Build agentic workflows using LangChain, LlamaIndex, n8n, Make, or whatever gets the job done best
- Integrate with OpenAI, Anthropic, and other model providers, and know when to use which
- Make sure agents are reliable, observable, and easy to debug when they go wrong
- Build evals so you know when an agent is performing well and when it has quietly started doing the wrong thing
WHAT WE ARE LOOKING FOR
- Minimum 1 year of experience building AI agents that run in production, not just demos or side projects
- Hands-on experience with LangChain or LlamaIndex for agent and RAG pipeline development
- Working knowledge of vector databases such as Pinecone, Weaviate, Chroma, or similar
- Strong experience with OpenAI and Anthropic APIs including function calling, tool use, and prompt engineering
- Comfortable building automation workflows in n8n, Make, or Zapier
- Strong Python skills, this is not optional
- You understand RAG deeply, chunking strategies, embedding models, retrieval quality, re-ranking, and how to debug when retrieval goes wrong
- You think in systems, not scripts
THE AI-FIRST REQUIREMENT
This one should go without saying for this role, but we will say it anyway. We are not looking for someone who is learning AI on the job. We want someone who has already shipped agents that real users or real systems depend on. Someone who has hit the hard problems, hallucinations, retrieval failures, broken tool calls, flaky orchestration, and knows how to solve them. If you have a portfolio of agents you have built, tell us about them. That is the most important thing you can show us.
NICE TO HAVE
- Experience with multi-agent frameworks like CrewAI, AutoGen, or similar
- Familiarity with voice AI systems, ASR, TTS, and conversation state, given DialNexa's core product
- Experience building internal tooling or developer platforms
- Comfort working across the DialNexa stack: Python, FastAPI, Postgres, Redis, Azure, GCP
WHAT SUCCESS LOOKS LIKE IN 6 MONTHS
- A live, maintained company knowledge base that agents across functions are actively querying
- At least one agent running in production across each major function, operations, marketing, sales, and engineering
- Measurable reduction in manual, repetitive work across the team as a direct result of what you have built
- A clear roadmap for the next set of agents and the infrastructure improvements needed to support them
WHY DIALNEXA
- You will not be building AI tooling to support a team of hundreds. You will be building the system that allows a small team to operate like a much larger one. That is a different and more interesting problem.
- Direct collaboration with founders on what gets built and why
- Full autonomy on architecture and tooling decisions
- Competitive compensation with ESOP discussion for the right person
- Cell phone reimbursement, health insurance, and paid time off
Click on Apply to know more.