About
I am a PhD student and Graduate Research Assistant at Louisiana State University (LSU), where my research sits at the intersection of computer vision, large language models, and multimodal AI. My current doctoral work investigates catastrophic forgetting in sequential fine-tuning of LLMs, and the application of LLM-based multi-agent systems to smart-grid management (AI Copilot for GE Vernova GridOS).
In parallel, I am the founder of Gradmate.ai — an AI-powered platform that streamlines graduate school applications through LLM orchestration, program matching, SOP generation, and IELTS practice. Prior to LSU, I was a Graduate Research Assistant at Alfred University, where I won 1st place (tied) in the Advizex AI Innovation Challenge with an AI sales bot projected to improve ROI by 25%.
My publication record spans IEEE conferences, ICPR, and arXiv, with work in facial attribute captioning, sign-language recognition, low-resolution license plate recognition, and toxic content classification. I have also released open datasets and fine-tuned models on HuggingFace that are actively used by the community.
Research Interests
Selected Projects
A face-swapping pipeline that disentangles facial identity from expression and appearance in StyleGAN's W-space. Introduces DeepFace neutral-emotion filtering to source only neutral-expression StyleGAN faces, paired with InsightFace gender validation for controlled attribute transfer. Key contribution: W-Space Refined blending outperforms Hybrid fusion (ID↔Src cosine similarity 0.359 vs. 0.232), enabled by convex-hull face masks from 106 landmarks, fixed Poisson blending with ROI boundary clamping, and pre-blend colour correction for perceptually natural composites.
FaceGemma enhances a Gemma-based vision-language model with an explicit facial attribute encoder, producing rich, identity-aware captions for portrait images. Trained on the custom FaceAttDB multilingual dataset (Zenodo), the model improves BLEU and CIDEr metrics over general-purpose captioners by conditioning generation on age, gender, emotion, and facial geometry.
A two-stage deep-learning pipeline for recognising Bangla license plates under real-world low-resolution conditions. A YOLO-based detector localises the plate region, followed by a CNN-LSTM sequence model for character recognition. Evaluated on a custom dataset captured from surveillance cameras, achieving state-of-the-art recognition accuracy for Bangla script.
An automatic portrait colorization framework that leverages identity-aware features extracted from a pretrained VGG-Face network as conditioning signals for a CNN colorization decoder. By anchoring colour prediction to facial identity features, the model produces perceptually plausible skin-tone and hair-colour outputs that are consistent across different images of the same subject.
Founded and engineered Gradmate.ai, an end-to-end LLM-powered platform for graduate applicants. Core features include semantic program discovery (dense retrieval over university programme embeddings), multi-agent SOP/LOR generation via LangGraph, IELTS adaptive practice, and a community knowledge base. Stack: LangChain · LangGraph · Groq · DeepSeek · Next.js. Currently seeking seed investment to scale.
Doctoral thesis developing a multi-agent LLM system integrated with GE Vernova's GridOS platform for real-time power grid decision support. Agents handle fault diagnosis, load-balancing recommendations, and operator-alert summarisation using LangChain tool-calling over ADMS/AEMS APIs. Evaluated against operator response latency and decision-accuracy benchmarks.
Systematic empirical analysis of how LLMs degrade on previously learned NLP tasks after sequential domain fine-tuning. Benchmarks span GLUE and SuperGLUE tasks with controlled fine-tuning order experiments across Orca, Qwen, and other open-weight models. Explores elastic weight consolidation, LoRA merging, and replay strategies as mitigation techniques. Under review at ACL ARR 2024.
A real-time Bangla Sign Language finger-spelling system built on a custom-trained YOLOv8 object detector operating at 30+ FPS on commodity hardware. A curated dataset of 36 Bangla hand-gesture classes was collected under varied lighting and background conditions, achieving 94.3% mAP. Published at ICRIC 2023.
Publications
Under Review
IEEE
IEEE
IEEE
Experience
Present
- LangChain & LangGraph for multi-agent orchestration
- Groq and DeepSeek for low-latency LLM inference
- RAG-based program discovery and SOP generation pipelines
Present
- Sequential fine-tuning experiments on GLUE/SuperGLUE benchmarks
- LLM copilot system for GE Vernova GridOS (thesis)
Dec 2025
Jan 2024
Aug 2023
- Courses: AI, Computer Vision, Neural Networks, Pattern Recognition
- Supervised capstone projects in face recognition and object detection
Education
Present
Technical Skills
Python · C/C++ · MATLAB · JavaScript
OpenCV · YOLOv8 · CNNs · Detectron2 · DLIB
StyleGAN · DeepFace · InsightFace · Diffusers
PyTorch · TensorFlow · Keras · Scikit-learn
HuggingFace · LangChain · LangGraph · Transformers
Groq · DeepSeek · Ollama · vLLM
Docker · Git · Google Cloud · W&B · LaTeX
GE ADMS · AEMS · GridOS · Power Analytics