Sinjini Mitra

AI Researcher | ML Scientist | Creative Thinker

Solving real-world problems with AI in GenAI, healthcare, scientific discovery, and document intelligence.

📄 Resume   |   💻 GitHub   |   🔗 LinkedIn   |   ✉️ Email

🔬 Projects

---

Resume Screener AI

Semantic resume ranking engine using LLM-based embeddings and job description matching.
Tags: #LLM #NLP #SemanticSearch #FastAPI
Links: GitHub

---

Segmentation with GANs

Data-efficient segmentation pipeline using a fine-tuned StyleGAN model with a custom ToRGB module for stylized mask generation.
Tags: #GAN #ComputerVision #Segmentation #Low data
Links: GitHub

---

Histopathology Image Classification (Public)

CNN pipeline for cancer detection on PatchCamelyon histopathology dataset with GradCAM visualizations.
Tags: #ComputerVision #MedicalAI #GradCAM #PyTorch
Links: GitHub

---

Local LLM-based RAG System (Public)

Built a lightweight Retrieval-Augmented Generation system with local LLMs for document-based Q&A.
Tags: #RAG #LLM #FAISS #LangChain
Links: GitHub

---

Learning the Chart of Nuclear Isotopes using GNNs (Private Summary)

Graph neural network models for structure-aware prediction of nuclear cross-section data.
Tags: #GNN #ScientificML #HeterogeneousGraphs #PyTorchGeometric
Links: GitHub | arXiv

---

Mobility Prediction Using Graph Neural Networks (Private Summary)

Private Summary
Modeled spatiotemporal mobility patterns using graph-based deep learning techniques. Developed custom edge- and node-level feature transformations for predictive tasks.
Keywords: Mobility Prediction, Graph Neural Networks, Spatiotemporal Modeling.

---

Notes

• Some research projects are private due to collaboration agreements or unpublished status. Summaries have been provided to demonstrate technical contributions.
• For selected private work, detailed descriptions and results can be discussed upon request during interviews.