Applied ML Engineer · Open to Roles

Hey, I'm Parth
Tyagi.

B.Tech Mathematics & Computing · CUK

4 deployed AI apps · Rยฒ=0.995 · 90%+ accuracy. Full-stack ML engineer โ€” model training, Flask APIs, deployment. NLP · Recommendation Systems · Applied AI.

PythonScikit-LearnXGBoostNLPStreamlitNumPyPandas
GitHub Profile ๐Ÿ“„ Resume Get In Touch View Projects
parth@ml-lab: ~
Welcome to Parth's portfolio terminal!
Type help to explore.
โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
parth@ml-lab:~$
try: help · projects · resume · skills · contact · algos · github · clear
01 about

Who I Am

I'm a B.Tech student in Mathematics & Computing with a focused interest in Applied Machine Learning and AI system design. My work centres on building complete, production-ready ML pipelines โ€” from raw data and feature engineering through to model evaluation and live deployment.

I value clean code, honest model evaluation, and systems that actually work outside of notebooks. Every project I ship is a fully deployed Streamlit application with reproducible notebooks and professional documentation.

"Build things that work. Document things that last."

Currently building towards a career in ML Engineering, with hands-on experience across regression, NLP, and recommendation systems โ€” all deployed as real applications.

Email Me LinkedIn
Core Skills
Python · AdvancedScikit-LearnXGBoostNLP · TF-IDFPandas · NumPyStreamlitML PipelinesGridSearchCVCosine SimilarityMatplotlib · SeabornSQLGit · GitHubFeature EngineeringEDA
PythonXGBoostNLPStreamlitNumPyPandasSklearnGridSearchCVSeabornGitREST APIJoblibEDAFeature Eng. PythonXGBoostNLPStreamlitNumPyPandasSklearnGridSearchCVSeabornGitREST APIJoblibEDAFeature Eng.
02 projects

Featured Work

AI Fitness Intelligence
Body fat prediction · Workout planner · XGBoost vs Linear Regression. Rยฒ=0.995 · MAE=0.21%
XGBoostStreamlitSklearn
Gym Intelligence
AI Fitness API
Full-stack Flask app · Google OAuth · PostgreSQL · 12+ biometric inputs · Dark neumorphic UI · Live on Render
FlaskREST APIPostgreSQLThree.js
Live Demo โ†—
AI Job Skill Analyzer
NLP skill extraction · Gap analysis · Deployed Streamlit app matching profiles to roles
NLPStreamlitPython
Skill Gap AI
Movie Recommender
Content-based filtering via TF-IDF + cosine similarity on TMDB metadata. Live top-N recommendations.
TF-IDFCosine SimStreamlit
Movie Rec System
Women's Cloth Sentiment
NLP-based positive/negative review classification on women's fashion data. Retail-ready prediction pipeline.
NLPSentimentClassification
Sentiment NLP
Heart Disease Prediction
Clinical ML classifier predicting cardiac risk from patient features. Logistic Regression · Random Forest · Feature importance analysis.
ClassificationSklearnEDA
Cardiac Risk AI
03 algorithms

Algos From Scratch

Hand-coded ML algorithms using only NumPy โ€” no sklearn shortcuts. Pure math, pure understanding.

// 01
Simple Linear Regression
Single feature predictor built from first principles using gradient descent and the normal equation. Manual weight update loop with convergence tracking.
ŷ = w₀ + w₁x  ·  J = (1/2m)Σ(ŷ−y)²
View Repo →
// 02
Multi-Linear Regression
Extended to n features using matrix multiplication. Implements normal equation and batch gradient descent from scratch with NumPy only.
θ = (XᵀX)⁻¹Xᵀy
View Repo →
// 03
Perceptron
The foundational neural unit. Binary classifier with manual weight update rule โ€” the building block that started deep learning. Step activation function.
ŷ = step(Σwᵢxᵢ + b)
View Repo →
// 04
Gradient Descent
Batch, Mini-batch & Stochastic variants coded manually. Loss per epoch tracked, convergence visualized, learning rate effects compared.
w := w − α · ∂J/∂w
View Repo →
// 05
Logistic Regression
Binary classification using sigmoid activation. Cross-entropy loss, decision boundary visualization, manual forward + backward pass โ€” no sklearn.
σ(z) = 1 / (1 + e⁻ᴿ)
View Repo →
04 contact

Get In Touch

Let's Build
Together.

Open to ML Engineering roles, research collaborations, and interesting AI projects. Let's collaborate โ€” open to internships and product work.

๐Ÿ“ง parthtyagi3389@gmail.com
๐Ÿ“ Central University of Karnataka, India
✅ Message sent! I'll reply within 24 hours.