AI product development · Dhaka, Bangladesh
Build the product around the model
Dhaka-based AI engineering for LLM features, retrieval systems and agent backends, designed as parts of a working product rather than isolated demos.
What I can help build
LLM product features
Integrate language models into an application with bounded inputs, clear fallback behavior and a user experience that explains what the system can do.
Retrieval and semantic search
Build around embeddings, retrieval and application data when a product needs answers grounded in its own information.
Agent backends
Design APIs and task flows for agent-oriented products, with explicit lifecycle states and observable outcomes.
Full-stack delivery
Connect the model-facing backend to a responsive product interface, from API and persistence to the screen.
How I approach delivery
Start by defining the user, the task and what a useful result looks like. Then choose the smallest suitable model and application design, test the failure paths, and measure behavior before making performance claims.
Bangla AI and product-specific language support
AI products for Bangladesh may need more than a translated interface. Route52 is a Bangla-native assistant project with persistent conversations and agent-oriented workflows. Its case study shows how I approach language-aware product design alongside the underlying AI and application engineering.
Read the Route52 case studyDiscuss a project
Send the product context, the users it serves and the part you need help with. We can agree on scope before work begins.
Contact