Darshan Joshi

AI Engineer · Computer Vision · NLP · Cybersecurity

joshidarshan193@gmail.com+1 947 240 5099Southfield, MI, USAlinkedin.com/in/darshanjoshi05github.com/darshanjoshi05

Summary

AI Engineer with hands-on research experience in computer vision, NLP, and cybersecurity. Research Assistant at Lawrence Technological University developing YOLOv8-based classroom behavior detection with Grad-CAM explainability. 4 research publications spanning cybersecurity infrastructure, AI behavior detection, seismic deep learning (CNN-LSTM), and NLP document intelligence. EC-Council certified across 5 cybersecurity domains. Seeking AI/ML, backend, or software engineering roles.

Education

Master of Science, Computer Science Engineering

Lawrence Technological University · Southfield, MI, USA

GPA: 3.75 / 4.0 (US Scale)

2024 – 2025

Bachelor of Engineering, Computer Science Engineering

Sree Dattha Institute of Engineering & Science · Hyderabad, India

GPA: 6.5 / 10 (Indian Scale)

2019 – 2023

Experience

Research Assistant

Lawrence Technological University · Southfield, MI

Dec 2024 – Present

·Led end-to-end development of AI-based classroom behavior detection system using YOLOv8 with measurable Precision, Recall, and mAP@50 performance metrics.
·Designed and curated structured multi-class datasets (11+ behavioral classes), enforcing annotation quality under limited hardware resources.
·Integrated Grad-CAM explainability techniques to enhance model interpretability and support responsible AI practices.
·Conducted iterative hyperparameter optimization and documented reproducible YAML-based training workflows.
·Collaborated with faculty advisors and presented findings in structured research updates.

Projects

Resume Intelligence Engine

FastAPI · Python · NLP · ReportLab · python-docx

2026 – Present

·Architected a modular FastAPI backend for dynamic resume and cover letter generation from job descriptions.
·Built NLP-driven JD parsing pipeline using regex and heuristic classification (87%+ suitability score).
·Developed scalable PDF/DOCX generation with ReportLab and python-docx.

Classroom Behavior Detection

YOLOv8 · PyTorch · Grad-CAM · OpenCV · YAML

Dec 2024 – Present

·Designed and trained YOLOv8 multi-class detection model (11+ behavioral classes) achieving mAP@50 = 0.91.
·Integrated Grad-CAM heatmap visualization for model explainability and responsible AI validation.

Buffer Overflow Attack Lab

C · GDB · x86 Assembly · ASLR · Linux

Jan – Jun 2024

·Analyzed stack smashing, heap overflow, and format string attacks in 32-bit Linux environments.
·Evaluated ASLR and stack canary mitigations using GDB; achieved root access via exploit injection.

Global Harvest Imports Database

SQL · MySQL · ERD · DFD · Schema Design

Jan – Jun 2024

·Designed a normalized relational database (3NF) from stakeholder requirements using ERDs and DFDs.
·Defined referential integrity constraints across 5+ department tables.

Publications

·Joshi, D. (2023). Design and Analysis of Cyber Security Infrastructure in Large Enterprises and Organisations. IJARESM. Published

·Joshi, D. (2025). Classroom Behavior Detection Using YOLOv8 and Explainable AI. IJSRST. Published

·Joshi, D. et al. (2025). Earthquake Prediction and Synthetic Seismogram Generation Using Hybrid CNN-LSTM Model. American Journal of Civil Engineering. Published

·Joshi, D. (2026). NLP-Driven Resume Tailoring: A Modular Approach to JD-Aware Career Document Generation. IJCST. In Preparation

Technical Skills

AI / ML: Python, YOLOv8, PyTorch, OpenCV, Grad-CAM, Scikit-Learn

NLP / Backend: FastAPI, NLP, Regex, ReportLab, python-docx, Jinja2, REST APIs

Cybersecurity: Buffer Overflow, GDB, ASLR, Network Defense, Digital Forensics

Languages: Python, C/C++, Java, SQL, HTML/CSS

Tools: Git, YAML, Matplotlib, MySQL, Linux, VS Code

Concepts: Object Detection, Explainable AI, Schema Design, API Design

Certifications

Network Defense Essentials (EC-Council)Ethical Hacking Essentials (EC-Council)Digital Forensics Essentials (EC-Council)Cisco Labs Crash Course (EC-Council)Website Hacking Techniques (EC-Council)