Hi, I'm Shanta Majumder 👋

Software Engineering Graduate · Data Science · AI Research · Open Source

I am a Software Engineering graduate specialising in Data Science from Daffodil International University, with a strong interest in building efficient and reliable AI systems.

My current research interests centre around:

• Natural Language Processing (NLP)
• Retrieval-Augmented Generation (RAG)
• Agentic AI and multi-step reasoning
• Small Language Models (SLMs)
• Resource-efficient and privacy-preserving AI
• Weather Forecasting
• Quantum Computing

I am particularly interested in designing AI systems that are not only accurate, but also efficient, verifiable, and practical in resource-constrained environments.

📄 I am the author of a preprint on an Agent-Verified RAG architecture using Small Language Models, currently under peer review.

My undergraduate thesis focused on developing an agent-verified Retrieval-Augmented Generation framework for resource-constrained environments.

Key areas of work included:

• Designing dual-retrieval and verification pipelines using LangGraph
• Exploring multi-step reasoning and agentic verification
• Working with Small Language Models (SLMs)
• Evaluating Context Relevance, Faithfulness and Semantic Similarity
• Benchmarking on MS MARCO, TriviaQA, HotpotQA and 2WikiMultiHop
• Analysing inference and pipeline latency

📄 Publication

An Agent-Verified System for Resource-Constrained Query-Focused Summarisation using Small Language Models

Shanta Majumder · Suprove Chandra Sarkar

Preprint: Research Square / ResearchGate Currently under review in SN Computer Science

I contribute to open-source machine learning infrastructure and documentation, with a focus on deep-learning-based time-series forecasting.

Current contributions include:

• Reproducible API usage examples and executable doctests
• Documentation for RecurrentNetwork, NHiTS and TiDE
• Minimal TimeSeriesDataSet pipelines
• Architecture-specific input_chunk_length and output_chunk_length constraints
• CI/CD and linting improvements using pre-commit and Ruff
• Testing with lightweight PyTorch Lightning configurations

🔗 Pull Requests:

• PR #2425 — RecurrentNetwork Example
• PR #2426 — NHiTS Example
• PR #2427 — TiDE Example

Currently exploring:

Agentic AI → RAG → Small Language Models → NLP → Efficient AI → Weather Forecasting → Quantum Computing

I am especially interested in research-oriented open-source projects where theoretical ideas can be translated into reproducible implementations.

B.Sc. in Software Engineering — Major in Data Science Daffodil International University, Bangladesh

CGPA: 3.72 / 4.00