Srinivas Allanki

• Built an AI-powered HR analytics assistant using the NeuralSeek platform to automate querying and analysis of employee performance data. • Configured AI agent workflows to retrieve, summarize, and analyze HR-related information from enterprise dataset. • Tested and validated AI-generated reports against sample datasets to improve reporting consistency and data accuracy. • Assisted in building retrieval-based analytics pipelines for internal organizational reporting and performance tracking.

• Engineered a real-time parking occupancy detection system using MobileNetSSDv3 and Deep SORT, reducing driver wait times by 40%. • Deployed a production-grade ML pipeline on Linux servers using Docker, ensuring high availability for video stream processing. • Designed a system that processes feeds from 23 RTMP security cameras using NVIDIA DeepStream for real-time detection. • Optimized and tested the system on NVIDIA Jetson Nano 4GB and 8GB devices to ensure smooth, high-performance operation.

• Developed advanced computer vision systems for autonomous robotic decision-making in competitive environments, improving target detection accuracy from 64% to 89%. • Designed real-time object recognition algorithms using TensorFlow Lite that enabled autonomous navigation with inference speeds below 45 ms per frame. • Deployed and tested optimized vision pipelines on embedded devices, achieving stable real-time performance at 30+ FPS. • Collaborated with cross-functional hardware and software teams to integrate vision systems with robotic components. • Designed custom 3D-printed robotic components and sensor mounts using Fusion 360 and Ultimaker Cura, enabling rapid prototyping and efficient hardware integration. • Developed functional sensor housings and robotic support components for real-world testing and embedded system deployment.

• Designed state machine-based architectures in LabVIEW for real-time image acquisition and sensor synchronization. • Developed computer vision models using the LabVIEW Analytics and Machine Learning Toolkit, achieving 82% accuracy in image classification tasks. • Integrated NI Smart Camera and NI-DAQ hardware, enabling real-time vision pipelines operating at 17 FPS. • Assisted in hardware-software integration and validation for industrial automation and computer vision systems. • Performed performance tuning and troubleshooting for real-time image processing applications, improving system reliability and stability.