class Devendra:
def __init__(self):
self.role = "CS Undergraduate — AI & ML Specialization"
self.university = "JECRC University, Jaipur"
self.cgpa = 8.68 # /10
self.grad_year = 2027
self.focus = ["Hybrid ML+LLM Systems", "NLP", "Anomaly Detection", "Agentic RAG"]
self.philosophy = "Ship it end-to-end — a model isn't done until someone can use it."
def currently_building(self):
return "AI that doesn't just predict — it explains, and knows when to ask for help."
me = Devendra()- 🔭 Currently building ThreatVision — an AI-powered cloud security monitoring platform
- 🧠 Obsessed with systems that combine classical ML + LLMs — fast and smart, not one or the other
- 📊 Ranked in a Kaggle anomaly-detection competition (CEIP-DS-JECRC) — imbalanced multivariate time-series, XGBoost, PR-AUC optimization
- 🌱 Always learning — currently deepening my grip on cloud-native security and agentic AI workflows
- ⚡ Fun fact: my fake-news detector automatically calls in a bigger LLM only when it's genuinely unsure — it knows its own limits
🛡️ CREDIBLEHybrid Fake News Detection Platform A dual-engine system: a 5-model ML ensemble handles most predictions instantly, and automatically hands off to an LLM only when confidence is low. Models: Logistic Regression · Random Forest · Naive Bayes · Gradient Boosting · Linear SVC
|
🕵️ ThreatVisionAI-Powered Cloud Security Monitoring Analyzes system logs in real time to flag brute-force attempts, unauthorized access, and irregular activity — before it becomes a breach. Approach: statistical anomaly detection over rigid rule-matching — catches what a static rulebook would miss.
|
|
Multi-Agent RAG Document Assistant A full-stack agentic system for semantic search and Q&A over PDFs — citation-backed, not just plausible-sounding.
|
Imbalanced Time-Series Anomaly Detection Multivariate sensor data, ~1% real anomalies — accuracy is a trap here, so the real work was elsewhere. Key insight: discrete log-encoded sensors needed custom binary indicator features — rolling averages alone hid the signal.
|
📚 Foundational repos (learning-in-public, click to expand)
| Repo | What's inside |
|---|---|
| Basics-Of-Python | Practical Python fundamentals — scripting → data science basics |
| Object-Oriented-Programming | Encapsulation, Inheritance, Polymorphism, Abstraction — clean examples |
| Functions | Python functions, decorators, closures, higher-order functions |
| File-and-Exception-Handling | Secure file I/O, context managers, production-grade error handling |
ML & Deep Learning
NLP, AI & Agents
Data & Analysis
Backend, Infra & Deployment
Tools
|
AI / Software Engineering Intern — IntersElite · Oct 2023 – Dec 2023 · Remote
|
B.Tech, CSE — AI & ML Specialization JECRC University, Jaipur, India · Jul 2023 – Jun 2027 · CGPA: 8.68 / 10
Open to AI/ML Engineer, Data Analyst, and Software Engineering roles & internships. If you're working on something that combines real data with real impact — let's talk.
"Building AI that doesn't just predict — it explains, and knows when to ask for help."