Thejaswin S
Software Engineer – AI focused on NLP, Computer Vision, Generative AI
Bengaluru, Karnataka, India
Joined January 2026
Summary
Published researcher in multimodal biometrics and explainable AI, with peer-reviewed conference and journal publications that focus on adaptive fusion of face and gait features for human identification. arxiv+3
Applied machine learning engineer who builds production GenAI and RAG systems—implemented Text2SQL and RAG chatbots, embedding finetuning for prior-art detection, and scalable microservice architectures for patent drafting and knowledge retrieval. thejaswin
Strong academic performer with formal training in AI/ML (B.Tech) and applied data science (PGP Diploma), recipient of academic and competition recognitions such as an IAPR student paper award and datathon placements. thejaswin+2
Full-stack ML tooling and deployment experience across frameworks and platforms (TensorFlow, PyTorch, LangChain, FastAPI, Docker, AWS/GCP), indicating capability to take models from research to integrated production services. thejaswin+1
Work
Education
Projects
Writing
Do we need to install tesseract seperately?
September 1, 2024Technical blog post discussing use and installation considerations for Tesseract OCR in practical Python workflows.
Exploring Fusion Techniques and Explainable AI on Adapt-FuseNet: Context-Adaptive Fusion of Face and Gait for Person Identification
January 1, 2024Paper on context-adaptive multimodal fusion of face and gait for person identification, exploring fusion techniques and explainable AI approaches for biometric identification.
Adapt-FuseNet: Context-aware Multimodal Adaptive Fusion of Face and Gait Features using Attention Techniques for Human Identification
January 1, 2023Conference/journal work presenting Adapt-FuseNet, an attention-based adaptive fusion strategy for combining face and gait features for robust human identification.
Mapping Distinct Source and Target Domains on Amazon Product Customer Critiques with Cross Domain Sentiment Analysis
January 1, 2022Conference paper on cross-domain sentiment classification applied to Amazon product reviews, detailing preprocessing, feature extraction, and cross-domain classifier evaluation.