kandyana venkata sai dinesh

Computer Science Engineering Student | AI/ML

Visakhapatnam, IndiaArtificial Intelligence and Software Development
1roles
26skills
1education

About

Computer Science engineering student with experience in machine learning research, computer vision, LLM applications, RAG systems, backend APIs, and AI evaluation.

Experience

ML Research Intern

IIT Kharagpur

May - July 2025 · India

Processed large-scale LiDAR point cloud datasets using Python-based ML pipelines; implemented and trained the RandLA-Net deep learning model for semantic segmentation of TLS point cloud data; achieved 99.6% classification accuracy in detecting tree stems; built automated pipelines for preprocessing, noise filtering, ground segmentation, and feature extraction; collaborated with researchers to validate AI-generated segmentation results and support reproducible experimentation.

Education

Guru Ghasidas Vishwavidyalaya, A Central University Bilaspur (C.G.)

B.Tech, Computer Science and Engineering

Dec. 2022 - June 2026

6.5/10

Skills

PythonJavaScriptSQLLLMsPrompt engineeringAI response evaluationModel evaluationRAGData AnnotationData validationDeep learningComputer visionFact verificationNode.jsExpress.jsREST APIsFastAPIMySQLOpenAI APILangChainHugging FaceGitDockerAWSLinuxVS Code

Projects

VisionOps AI – Real-Time Computer Vision Analytics Platform

Built a real-time people detection and tracking pipeline using YOLOv8 and OpenCV; designed backend APIs for AI detection outputs and structured analytics; developed a React dashboard for occupancy, visitor count, and time-series activity insights; containerized the full stack using Docker and Docker Compose.

AI Multi-Agent Workflow Platform

Engineered a modular multi-agent architecture with specialized LLM agents, a FastAPI orchestration layer, prompt pipelines, inference routing, output evaluation, and Dockerized agent services, APIs, and frontend components.

AI-Powered Research Assistant Platform

Built a retrieval-augmented generation assistant for large document collections; developed APIs for query processing, document retrieval, and LLM response generation; implemented document embeddings and vector search; evaluated responses for factual correctness, relevance, and completeness.