Build the Intelligence Infrastructure That Matters.
At Zunn, we don't build consumer chat toys. We solve high-stakes engineering problems across multimodal statutory RAG, deterministic tax computation, high-throughput ecommerce automation, and sub-300ms neural speech synthesis.
An Engineering Culture Designed for High Agency.
We operate with high ownership, small elite teams, and zero tolerance for superficial vanity metrics.
Extreme Technical Rigor
We solve hard, real-world computational problems: multimodal legal parsing, deterministic tax engines, real-time store mutations, and sub-300ms neural audio streaming. We hold our code to institutional mathematical standards.
Real Economic & Institutional Impact
Our software protects enterprise tax compliance under the Income Tax Act 2023, automates mission-critical commerce operations, and brings voice accessibility to over 250 million Bengali speakers.
High-Ownership Engineering Craft
We operate with minimal bureaucracy, direct technical ownership, and deep uninterrupted focus blocks. Every engineer has direct visibility into customer impact and platform performance.
Engineering, Research & Architecture Roles
All roles based at our Dhaka Engineering HQ or open to global remote contributors.
Senior Agentic RAG & Knowledge Systems Engineer
Role Overview
Design multimodal statutory ingestion pipelines, pgvector HNSW indexing, and LangGraph agentic reasoning over Bangladesh legal acts, SROs, and case law precedents.
Key Technical Requirements
- Extensive experience with Python, TypeScript, PostgreSQL, and vector indexing (pgvector, HNSW).
- Demonstrated expertise with Agentic RAG frameworks (LangGraph, LlamaIndex) and coordinate bounding-box citation rendering.
- Deep understanding of structured document extraction from degraded Bengali and English legal gazettes.
- Experience optimizing high-accuracy retrieval for domain-specific regulatory corpora.