Chenna Yaswanth

Machine Learning Enthusiast | Deep Learning | NLP | Computer Vision | Generative AI | Python

Rājahmundry, IndiaArtificial Intelligence and Machine Learning
38skills
6education
3credentials

About

I am a B.Tech student in Computer Science and Engineering (CSE) with a specialization in Machine Learning. While ML is my domain, I enjoy working on diverse problem-solving challenges in computing. My projects include suicide rate analysis, text generation, and text summarization, where I have applied techniques like Markov models, LSTMs, and deep learning. Beyond ML, I am keen on exploring broader areas in computer science, including software development, UI/UX, and data-driven solutions. I am open to learning, collaborating, and building innovative solutions.

Education

Lovely Professional University

B.tech, Computer Science and Engineering

Sep 2022 - Jul 2026

CGPA: 6.61

AAKASH INSTITUTE

Intermediate, M.P.C

May 2020 - Jun 2022

AAKASH INSTITUTE

12

[object Object]

Sri Chaitanya

10th

[object Object]

Sri Chaitanya Techno School

SSC

[object Object]

Mathrusri Junior College, Rajahmundry

Intermediate, MPC

Apr 2020 – Mar 2022

Percentage: 80.7%

Skills

Python (Programming Language)JavaC++Natural Language Processing (NLP)PythonRegressionClassificationClusteringFeature EngineeringModel EvaluationCNNsTransformersTransfer LearningFine-tuningTrOCRBARTflan-t5RAGLLMsMulti-Agent SystemsText SummarizationEmbeddingsSemantic SearchPyTorchscikit-learnHugging Face TransformersLangChainFAISSNumPyPandasFastAPIGradioGoogle ColabHugging Face SpacesHugging Face Model HubGitVS CodeLinux

Projects

Handwritten Text Recognition using Transformer-based OCR

Fine-tuned Microsoft TrOCR on 2,500 IAM handwriting samples, achieving CER below 7% and WER below 15%. Resolved an RGB/grayscale channel mismatch bug, trained a 1.3GB model on Colab GPU, hosted it on Hugging Face via Git LFS, and deployed a FastAPI and Gradio inference pipeline on Hugging Face Spaces.

AI Meeting and Dialogue Summarizer with RAG

Fine-tuned BART-large-CNN on 3,000 DialogSum samples and integrated FAISS retrieval. Engineered a Writer-Critic multi-agent pipeline using flan-t5 models for iterative answer refinement, then deployed PDF upload, question answering, and meeting summarization on Hugging Face Spaces via Gradio.

Multi-Agent Document Question Answering System (RAG)

Architected a LangChain-orchestrated Writer and Critic multi-agent RAG system with FAISS semantic search for document-grounded question answering. Deployed an interactive Gradio application on Hugging Face Spaces supporting multi-document ingestion and context-aware answers.