Data Scientist & Data Analyst
Hi, I am
Shariq Mukadam
My Intro
I'm Shariq Mukadam, a Computer Science student passionate about Data Science, Data Analytics.
My core focus is Artificial Intelligence & Machine Learning — building predictive models using classification, regression, and clustering, and exploring deep learning with TensorFlow and PyTorch. Equally central is Data Analytics — crafting dashboards in Power BI and Tableau, writing complex SQL, and turning raw data into business insights. Frontend is a complementary skill — clean, responsive interfaces with HTML, CSS, JS and Tailwind CSS.
I'm actively seeking internships, full time oppurtunities, and research collaborations where I can contribute real value and grow alongside a driven team.
Exploring insights through Data Science & Analytics
Multi-modal ML system predicting e-commerce return risk trained on 66M Amazon reviews across 611K products. Fuses CLIP image embeddings, sentence-transformer review encodings, and tabular rating signals via a custom PyTorch attention fusion layer into LightGBM, achieving AUC-ROC 0.83. Per-prediction SHAP explainability is served through a FastAPI endpoint with a full React 18 seller dashboard.
End-to-end AI job market intelligence platform scraping 300+ live JDs across LinkedIn, Naukri & Internshala with Groq LLM skill extraction and K-Means clustering by company tier.Features a genuine LangGraph ReAct agent with multi-turn memory that delivers skill graph and market cluster via Streamlit.
Something exciting is in the works. Stay tuned.
End-to-end data analytics project tracking India's UPI payment revolution across 77 months (Nov 2019–Mar 2026) using 13 verified RBI & NPCI government CSVs. Built a full pipeline: data cleaning → feature engineering → PostgreSQL → 5 SQL queries → 7 EDA charts → 3-page interactive Power BI dashboard. Real government data. No Kaggle.
ETL pipeline on 6 years of MoRTH government CSVs + PDF extraction → PostgreSQL + 5 SQL queries + Power BI dashboard; normalized fatality rates revealed Bihar (0.81) is 10x more dangerous per crash than Kerala (0.08); Prophet model forecasts India crossing 2 lakh fatalities by 2027; over-speeding confirmed as cause of 68.1% of deaths vs drunk driving at 2.1% for drunk driving.
A real-time, interactive dashboard tracking the Analytics & AI job market across India. Features a custom dark-glassmorphism UI, multivariate market analysis, and live data aggregation from LinkedIn, Naukri, and Internshala.