praveena = {
"name" : "Praveena Simhadri",
"education" : "Dual Degree (B.Tech Hons. + M.Tech) | CSE | IIT Kharagpur",
"role" : ["Software Engineer @ BlueFlag Security", "Ex-Full Stack Intern @ Zolve"],
"research" : ["NLP & Bias in LLMs", "Graph Neural Networks", "Computer Vision"],
"passion" : "Building AI systems that are interpretable, fair, and impactful",
"portfolio" : "https://simhadripraveena.bolt.host/",
"status" : "Open to research collaborations & ML/AI roles 🚀"
}| # | Paper | Venue | Status |
|---|---|---|---|
| 1 | LinguisticRedline: Uncovering Racial Bias in LLM Perceptions of Urban Crime Risk | NLPercep'26 — Workshop on Centering Social Perception in NLP, co-located with ICWSM'26, Los Angeles, USA | ✅ Accepted |
| 2 | Robustness Evaluation of Deep Learning-Based Plant Disease Classification under Adversarial and Physical Perturbations | NetCrypt-26 — 4th International Conference on Networks and Cryptology | 🔄 Under Review |
| Competition | Organizer | Project | Status |
|---|---|---|---|
| Gemma 4 Good Hackathon | Kaggle × Google DeepMind | TruthLens — AI-powered misinformation detection | 🔄 Submitted · Results Pending (May 2026) |
| 🧠 | Topic |
|---|---|
| 🔤 | NLP & Fairness — Bias detection in language models, social perception in AI |
| 🕸️ | Graph Neural Networks — Spatio-temporal learning, flood & climate modeling |
| 👁️ | Computer Vision — Deep learning-based recognition & classification systems |
| 🌍 | AI for Social Good — Climate, security, interpretability, ethics |
| 🔹 Project | 💡 Description | 🛠️ Stack |
|---|---|---|
| LinguisticRedline | NLP research probing racial bias in LLM crime-risk perceptions — published at ICWSM'26 | Python, HuggingFace, LLMs |
| Person Identification System | End-to-end face recognition pipeline with deep learning embeddings + geometric landmarks, real-time inference API | PyTorch, FastAPI, Django |
| Flood Graph (IIT KGP) | Spatio-temporal GNN framework for rainfall-to-inflow prediction; spatial clustering and graph-based event analysis | PyTorch Geometric, GNN |
| TruthLens | Misinformation detection and media analysis tool | Python, NLP |
| Plant Disease Classifier | Adversarially robust deep learning model for plant disease detection under physical perturbations | TensorFlow, CNN |
| House Price Predictor | XGBoost regression with full-stack deployment — interactive dashboard for real-time predictions | XGBoost, FastAPI, Django |
| Emotion Recognition App | CNN + TFLite mobile app for real-time emotion detection, 82% accuracy on Android | TFLite, CNN, Android |
- Resolved production & development tenant issues across backend services and data pipelines
- Built a Neo4j dependency graph to model complex service relationships; normalized MongoDB datasets
- Converted research notebooks into production ClickHouse analytics pipelines
- Developed a user hierarchy modeling system — later showcased by CEO to enterprise clients
- Built an internal customer data platform (Django + SQL) powering secure and scalable user insights dashboards
Languages
ML / DL
Data & Infrastructure
Software & DevOps
- 🐍 Python for Machine Learning & Data Science Masterclass — Udemy
- 🌍 MEGA Machine Learning in GIS & Remote Sensing — Udemy
- 🗄️ SQL, NoSQL, Big Data and Hadoop — Udemy
- 🎨 Introduction to Animation Fundamentals — Simplilearn
- 📊 Accenture North America Strategy Consulting Virtual Experience
- 🏦 JPMorgan Chase Commercial Banking Virtual Experience
"Research that challenges assumptions, code that scales with purpose, systems that serve people — not just benchmarks."
⭐ Open to research collaborations, ML engineering roles, and impactful AI projects. Let's build something meaningful.