Kaushik M

Final-year B.E. CSE student · AI Security & Generative AI Engineering

Chennai, Tamil Nadu, India

@kaushik_m1

LinkedIn GitHub

0 followers

🎓 Bachelor of Engineering (B.E.) in Computer Science and Engineering at Sathyabama Institute of Science and Technology · Graduating 2027

Follow Message

About

Final-year Computer Science and Engineering student with hands-on experience in AI, machine learning, and Generative AI across data collection, model development, evaluation, and deployment. Built LLM applications, multi-agent systems, and quantized model benchmarks, with experience translating technical findings into clear documentation and recommendations.

Experience

Software Engineer Intern

VCIDEX Solutions Pvt. Ltd. · Chennai

Developed backend components in Python for an AI-powered Product Analyzer application as part of a three-member engineering team, working within an existing codebase to add new functionality. Collaborated with internal team members to gather requirements, design, integrate, and test AI-driven analysis features, contributing to reliable end-to-end feature delivery.

Education

Bachelor of Engineering (B.E.) in Computer Science and Engineering

Sathyabama Institute of Science and Technology

Computer Science and Engineering · 2027

Skills

<3 months

SQL Oop Python Pandas PyTorch Testing Statistics TensorFlow Benchmarking Advanced DBMS Data Analytics Responsible AI Data validation Data Collection Machine Learning Prompt Engineering Model quantization Supervised Learning Multi-Agent Systems Unsupervised Learning Large Language Models Model Context Protocol Artificial intelligence Data Structures and Algorithms Retrieval Augmentation Generation

Tools / apps / platforms

<3 months

Git GitHub Ollama PyCharm ChromaDB Claude Code Google Cloud Platform Amazon Web Services AWS Grafana Labs Grafana Cloud Oracle Cloud Infrastructure

Projects

ARM Adaptive Guardrail - Quantized LLM Benchmarking & Middleware

Python, Ollama, Qwen2.5 (Quantized), ChromaDB, AWS Graviton

Trained and benchmarked a quantized Qwen2.5 language model on Arm64 architecture, improving throughput by 34% while validating output quality. Built adaptive middleware selecting among four model configurations by latency budget, reducing defect rates from 42.9% to 0% through iterative model tuning.
LLM Red Team Lab - Automated AI Security & Red-Teaming Framework

Python, Streamlit, Groq API, Cerebras API, ChromaDB, Claude Code, Pycharm

Engineered an automated red-teaming pipeline to systematically evaluate RAG-based chatbots against prompt injection attacks, including document-embedded and poisoned knowledge-base payloads, using ChromaDB for retrieval-layer testing. Designed and benchmarked five defense configurations across three LLMs, reducing confidential data leak rate by 93% while maintaining a 0% false-refusal rate under adversarial-exposure usability testing.

Courses & certifications

Generative AI Explained, Agentic AI Explained · NVIDIA
Generative AI Leader, AI Essentials, Prompting Essentials · Google
AI Fluency: Framework & Foundations, Claude Code 101 · Anthropic
Learn SQL Basics for Data Science Specialization · UC Davis
Generative AI Engineering Professional Certificate · IBM
Cloud Infrastructure (OCI) Certification · Oracle

Explore the Broxer community

💬 Q&A Ask about roles, salaries and interviews — and help others. 👥 Circles Join discussions on skills, companies and industries. 📰 Feed Follow updates and insights from the community.