Bonu Sasank

I am an aspiring AI/ML engineer holding a B.Tech in Electronics and Communication Engineering from SRM University (2025). Practical experience in developing end-to-end machine learning solutions, such as LLM-based workflows, YOLOv8-based SAR detection, and NLP pipelines for detecting fake news. Experienced in Python, TensorFlow, scikit-learn, and AWS tools. Enthusiastic about addressing real-world issues with generative AI, RAG pipelines, and contemporary MLOps workflows.

• Telugu
• Hindi
• English

Familiar with emerging technologies and tools, including

• LangGraph, Elastic, Kubernetes, Terraform, Ansible, and AIOps
• Basic understanding of observability practices in ML systems
• Actively exploring advanced DevOps and system integration concepts through self-paced labs and GitHub-based projects

Target detection and classification from SAR images

Feb 2025 – May 2025|Python, YOLOv8, and autoencoders

• Obtained approximately 18% improvement in target classification accuracy on SAR images by combining autoencoder-based denoising with YOLOv8, eliminated noise interference by more than 30% through SSIM improvement, enhancing detection accuracy in real-time environments, and facilitated a sub-100 ms latency-enabled detection pipeline for military-grade use cases

Satellite image denoising using machine learning

Aug 2024 – Nov 2024 | Python, scikit-learn, PSNR/SSIM, NLM

• Improved NLM with ML methods for better PSNR by 4.7 dB and SSIM by 12%
• Denoised more than 1,000 satellite images for better analysis and segmentation
• Enhanced downstream model accuracy by approximately 15% in land-use classification

Fake news detection using NLP

Jul 2023 – Sep 2023 | Python, scikit-learn, NLP

• Constructed a binary classifier with TF-IDF and logistic regression, 93.5% accuracy on a dataset of over 10,000 articles
• Applied lemmatization, removal of stop words, and n-gram analysis to increase relevance
• Validated model using 5-fold cross-validation with less than 2% accuracy variance