Goutham

SANDITI GOUTHAM REDDY Email-id:-Linkedin: https://www.linkedin.com/in/goutham-sanditi-38194b1b0 Mobile No.:- IIT Mandi EDUCATION • Indian Institute Of Technology, Mandi B.Tech - Computer Science and Engineering 2020 - 2024 Mandi, Himachal Pradesh TECHNICAL SKILLS • Programming Languages: Python, C++, Java, C#, JavaScript, SQL • AI/ML Frameworks: TensorFlow, PyTorch, Keras, Scikit-learn, Large Language Models (LLMs), U-Net, GANs, Quantum ML (QCNN) • Developer Tools, Databases & Systems: .NET Framework, Roslyn APIs, d3.js, Workato Automation, MySQL, MongoDB, PostgreSQL, Unix/Linux, Git, TFS WORK EXPERIENCE • EMA Unlimited, Inc July 2024 - Present AI Application Engineer ◦ Built Agent Assist’s AI pipeline, a production-ready system assisting customer support agents with end-to-end workflows including categorization, summarization, response generation, and validation. ◦ Integrated EMA with client ticketing platforms by compiling domain-specific business logic, enabling automated actions (tagging, updating fields, snoozing tickets) and achieving 97%+ response accuracy. ◦ Developed a computer-use PoC to extract structured insights from external websites without APIs, applying agentic AI to unstructured environments. ◦ Created an evaluation tool leveraging LLMs-as-judges for automated product QA, self-correction, and improvement suggestions, streamlining evaluation across customers. ◦ Designed and deployed multiple agentic AI pipelines for client PoCs, showcasing advanced prompt engineering, LLM orchestration, and pipeline optimization. ◦ Delivered scalable AI/ML solutions that improved automation and customer experience for enterprise clients. • Siemens Technology and Services Private Limited Jan 2023 - July 2023 Software Developer (Intern) ◦ Developed Reference Analyser, a tool that identifies and removes unused references in C# projects by analyzing dependencies through Reflection, Roslyn APIs, and .csproj file parsing, resulting in a 46% improvement in build time for a tested project. ◦ Implemented visual reference graphs using d3.js to provide insights into project structure and dependencies, aiding in optimization efforts. MAJOR PROJECTS • VineNet Apr 2024 Yamaha Hackathon ◦ Engineered a U-Net deep learning model for precise segmentation of grape bunches in high-resolution vineyard images, enhancing vineyard management with pixel-level accuracy. ◦ Optimized training with TensorFlow using Binary Cross-Entropy loss, Adam optimizer, and essential callbacks, achieving 86% Precision, 81% Recall, and 71% Mean IoU. • CT Scan Reconstruction Apr 2024 - May 2024 ◦ Created a denoising pipeline using Content-Noise Complementary Learning (CNCL) strategy with UNet and GAN framework, which demonstrated significant improvements in visual quality and quantitative metrics for denoising CT (DICOM) datasets RELEVANT COURSES • Deep Learning and its Application, Machine learning, Linear Algebra, Data Structures and Algorithms,Database Management System, Operation Systems, Probability and Statistics, Matrix Computations for DataScience.