← Portfolio
Shivansh Mishra
Professional Summary
Pre-final year B.Tech CSE student specializing in Machine Learning Engineering and MLOps. Strong background in pipeline design, Scikit-learn modeling, and containerized API serving. Experienced in hyperparameter tuning (GridSearchCV), cross-validation protocols, and model evaluation metrics (RMSE, F1, R²). Proven track record of architecting reproducible workflows and launching 10+ live predictive microservices. Passionate about deploying, monitoring, and scaling production-grade ML systems.
Education
Babu Banarasi Das University, Lucknow
2024 – Present
B.Tech in Computer Science & Engineering (Specialization: Cloud Computing & Machine Learning)
  • Academic Standings: CGPA: 8.0+ / 10. Relevant Coursework: Machine Learning Systems, Data Structures & Algorithms, Database Management (DBMS), Systems Design, Cloud Orchestration, Python.
Technical Skills
ML Modeling & Validation:
Scikit-learn GridSearchCV Cross-Validation Regression & Classification Residual Analytics
Data Engineering:
Pandas NumPy Feature Engineering Imbalance Handling
MLOps & Deployment:
Docker FastAPI Streamlit Git GitHub Actions Vercel Render
Core Languages & Sys:
Python SQL REST API Design Relational DBs
Projects
House Price Inference System — Modular ML App [Live] [GitHub]
2024 – 2025
Python, Scikit-learn, Pandas, Streamlit, Vercel, REST API Serving
  • Architected a modular Scikit-learn modeling pipeline integrating data cleaning, automated feature scaling, and estimator fitting.
  • Evaluated model algorithms (Linear vs. Ridge Regression) using RMSE and R² metrics to systematically select the highest-performing model.
  • Wrapped the inference model in a Streamlit container, deploying to Vercel with integrated automated validation checks.
Car Price Optimization — GridSearchCV Pipeline [Live]
2024 – 2025
Python, Scikit-learn, GridSearchCV, 5-Fold Cross-Validation, Pipelines
  • Engineered automated data preprocessing pipelines using custom classes for mileage decay and categorical target encoding.
  • Executed extensive hyperparameter optimization via GridSearchCV with 5-fold cross-validation, decreasing validation loss variance by 12%.
  • Configured reproducible artifact serialization, exporting trained pipeline states to ensure consistent out-of-sample inference.
Experience & Open Source Contributions
ML Engineering & Open Source Contributor
2024 – Present
  • Maintained 10+ public GitHub repositories demonstrating PEP8 coding standards, detailed workflow documentations, and reproducible configurations.
  • Published analytical datasets and model training logs on Kaggle, verifying pipeline inputs and validation notebook reproducibility.
  • Participated in Coderush 2.0 Hackathon (2025), creating optimized algorithms to resolve numeric constraints under strict time limits.