AI Engineer with 2+ years of research & industry experience turning LLMs, vision models and multimodal pipelines into production systems — currently engineering applied AI at eQOURSE.
Custom chatbots, RAG knowledge systems, LLM fine-tuning and prompt-engineered pipelines — grounded in your data, deployed on your infrastructure.
Multi-agent workflows, intelligent data pipelines and business-process automation — the kind that cut manual workflows by ~60% at Reality AI.
Object detection, tracking, segmentation, surveillance and video analytics — real-time systems proven at 97.47% production accuracy.
97.47% ACC / 100% RECALLHybrid CV + ML pipeline for real-time parking occupancy — 7.47% above industry benchmark — plus custom-YOLO human fall detection with live alerting dashboard.
<10S ZERO-SHOT CLONEReal-time zero-shot text-to-speech on a 0.5B-parameter Llama backbone — clones any voice from seconds of audio with emotion control, Streamlit UI and Python API.
4-MODEL MULTIMODALImage classification, speech-to-text severity analysis, U-Net satellite flood segmentation and a Gemini meta-model — unified for government disaster response.