Work arrangement: Full-time, remote within Bangladesh
Working hours: Sunday–Thursday, with at least four working hours overlapping with Saudi Arabia business hours
Salary: BDT 60,000–120,000 per month, based on demonstrated ability
Initial probation: Three months, paid
Start date: As soon as mutually agreed
About the company
We are building a private enterprise AI-agent platform that runs on customers’ own servers. The platform will allow employees to search internal documents, receive answers with citations, draft business content and use approved company tools without sending confidential company information to public AI services.
This is an early-stage product role. You will directly influence the architecture and development standards.
What you will do
• Build Python backend services using FastAPI.
• Integrate and serve locally deployed language models.
• Build retrieval-augmented generation pipelines.
• Process PDF, Word, Excel and text documents.
• Implement embeddings, vector search, metadata and reranking.
• Build controlled agent workflows and tool-calling services.
• Implement user permissions, audit logs and human-approval controls.
• Work with PostgreSQL and vector-search technology.
• Package the application using Docker.
• Write automated tests and technical documentation.
• Review junior-engineer contributions.
• Work with a part-time senior architect on production readiness.
Required experience
• Two to four years of professional software-development experience.
• Strong practical Python ability.
• Experience with FastAPI, Django REST Framework or Flask.
• Experience with PostgreSQL and SQL.
• Experience designing or consuming REST APIs.
• Working knowledge of Git, Linux and Docker.
• Experience writing automated tests.
• At least one completed LLM, RAG, semantic-search or document-intelligence project.
• Good written and spoken English.
Preferred experience
• Ollama, vLLM or llama.cpp.
• LangGraph, LangChain, LlamaIndex or similar.
• pgvector, Qdrant, Milvus or Weaviate.
• OpenSearch or Elasticsearch.
• Document parsing and OCR pipelines.
• Authentication and role-based permissions.
• Background jobs and message queues.
• Monitoring and structured logging.
A university degree is helpful but is not required. Demonstrated work is more important.
What you must submit
1. Your CV.
2. Your expected monthly salary in BDT.
3. Your earliest available start date.
4. Your GitHub, GitLab or portfolio link.
5. A short description of one relevant system you personally built.
6. A clear explanation of which parts you personally implemented.
7. Your computer specifications and internet reliability.
8. Your English availability for a live technical interview.
Selection process
1. CV and portfolio review.
2. Twenty-minute communication interview.
3. Sixty-minute live technical discussion.
4. Small paid practical assignment.
5. Reference and identity verification.
6. Three-month paid probation.
The paid task will involve building a small local document-question-answering API with source references. It will not require building our commercial product.
Applications that contain only certificates without demonstrable projects will not be prioritized.
Kindly share CV on mail address: [email protected] / [email protected]