About the role
Much of what TECFIQ sells runs on applied AI: Captura reads scanned paper and PDFs into structured data, XDOCS AI DMS and Ask The Archive answer questions from a customer's own documents, and ProHRM screens candidates with AI. As an AI / Machine Learning Engineer you will take those capabilities from experiment to dependable production features.
What you'll do
- Design, train and evaluate models and pipelines for OCR, document understanding and data extraction on Bangla and English documents
- Build and improve retrieval-augmented generation (RAG) pipelines and LLM-based assistants that answer from customers' own data
- Take models from notebook to production: package, deploy, monitor and iterate behind reliable APIs
- Create evaluation sets and metrics so accuracy, latency and cost can be tracked as products change
- Work with product engineers to ship AI features into TECFIQ products
- Handle messy real-world inputs such as low-quality scans, handwriting and mixed-language forms
- Keep up with new models and tooling, and judge which are worth adopting
What we're looking for
- Bachelor's degree in Computer Science, Engineering, Data Science or a related field, or equivalent practical experience
- Strong Python skills and experience with PyTorch or TensorFlow
- Practical experience in NLP or computer vision — OCR, document layout analysis or LLM-based systems
- Experience building RAG pipelines, embeddings and vector search
- Comfort with REST APIs, Docker, Git and Linux deployments
- Ability to explain model trade-offs clearly to non-ML teammates
- Good written English
Nice to have
- Bangla NLP or Bangla OCR experience
- Fine-tuning and prompt engineering of large language models
- Experience with FastAPI or similar frameworks
- Deploying models on cloud GPUs
- Experience with banking, finance or HR document workflows