Hi! I'm Harsh Wasnik, a Machine Learning Engineer with experience in building scalable AI systems across healthcare, education, and telecom.
- 🎓 MS in Information Science, University of Arizona (4.0 GPA, Distinguished Graduate Scholar)
- 🤖 Focus Areas: LLMs, Deep Learning, Graph Neural Networks, ML Ops
- 🧾 Published author @ IEEE
- 🚀 I use AI to solve real-world problems with measurable impact
| 🔍 Context-Aware Systems | 📦 MLOps Pipelines | 🧠 Deep Learning |
|---|---|---|
| LLMs + KGs for personalization | Airflow, MLflow, Docker | CNNs, LSTMs, Transformers |
| 📊 Predictive Analytics | ☁️ Cloud Platforms | 🗃️ Vector Databases |
|---|---|---|
| A/B Testing, Scoring Models | AWS (S3, EC2, SageMaker), GCP, Azure | FAISS, LangChain, Semantic Search |
- Built an end-to-end MLOps pipeline for real-time phishing detection using a network security dataset
- Modular architecture with YAML configs, CI/CD via GitHub Actions, and MLflow experiment tracking via DagsHub
- 🐳 Dockerized and deployed the system to AWS EC2 using ECR for scalable production use
- 🧪 Trained and compared models: Random Forest, AdaBoost, Gradient Boosting, Logistic Regression, Decision Trees
- 🔗 GitHub | DagsHub | Live App
- HIPAA-compliant system that generates patient-specific treatment recommendations using EHR, clinical notes, and medication data
- Built with Databricks Spark, LLMs, and Knowledge Graphs
- 🏥 Achieved 90% accuracy in diagnosis prediction
- Published at IEEE ICCPCT 2025
- Integrated GPT-4 with semantic knowledge graphs for personalized learning
- 📈 Improved student engagement by 25%, retention by 18%
- Predicted standardized test scores (math, reading, writing) with >90% accuracy
- Inference latency <1 second per profile across 100+ test cases
- 🔗Live App
- Deployed Graph Neural Networks to optimize telecom traffic from major clients like Amazon/Facebook
- Resulted in $80K in revenue through improved network efficiency
Languages: Python, R, Bash
ML Frameworks: PyTorch, TensorFlow, Scikit-learn, XGBoost
Tools: Apache Airflow, MLflow, Docker, Streamlit, HuggingFace, LangChain
Data: PySpark, PostgreSQL, MongoDB, Snowflake, Redshift
Cloud: AWS (S3, SageMaker, Lambda), GCP, Azure
Viz: Power BI, Tableau, Matplotlib, Plotly, Seaborn
Certs: Microsoft Generative AI, UIUC Analytics, Tata Excellence Award, Student Award, Distinguished Scholar
- 🌐 Portfolio: https://www.datascienceportfol.io/wasnikh0
- 🔗 LinkedIn: linkedin.com/in/harsh-wasnik
- Website: Personal Website
🧠 “Crafting Intelligent Solutions with AI”