Prateek Goswami

I’m a Computer Science & AI undergraduate with hands-on experience in NLP, deep learning, and end-to-end ML pipelines. I’ve focused on turning raw data into reliable, deployed solutions—especially where model decisions must be understandable and measurable.

As an AI / ML Engineer Intern at Patidar Agriculture, I built and deployed an NLP-powered chatbot to handle product queries and provide machinery recommendations. I also developed computer vision models for agricultural disease detection, using feature engineering, training, and evaluation with Python, OpenCV, and Scikit-learn, and integrated end-to-end ML into a live production website.

In my projects, I keep pushing toward real-world usefulness: I built an AI Outreach Agent that analyses company websites, matches opportunities to resumes, and generates personalised emails with hallucination detection and quality-scoring. I also created multimodal systems like a VQA model (ResNet/ViT + BERT), a real-time anti-cheating proctoring pipeline (YOLOv5, Mediapipe, TensorFlow), and a churn prediction pipeline on Databricks using PySpark ML and MLflow.

I enjoy the full lifecycle—from data cleaning and EDA to deployment-ready pipelines—while learning quickly and documenting clearly. Beyond building models, I lead with impact: as Organising Head for a college hackathon (200+ volunteers) and a Teaching Assistant for C programming (60+ students), I’ve strengthened teamwork, mentoring, and execution under real deadlines.