Hi, I'm Sunny
Data Analyst. AI/ML Data Associate. Occasionally sleeping.
Turning messy data into decisions — dashboards, ML pipelines, and labeled datasets that actually ship.
Open for full-time roles and data collabs.
→ Tula's Institute, Dehradun
About
I'm Sunny Kumar — a Data Analyst & AI enthusiast pursuing MCA (Final Year, 2026 Batch) at Tula's Institute, Dehradun.
I work with multi-modal data — text, speech, image, audio, video — labeling, cleaning, and shipping datasets that ML systems can actually learn from. My sweet spot is where analytics, NLP, and MLOps meet: Power BI dashboards, LLM-assisted annotation, and end-to-end forecasting stacks.
Fluent in Hindi & English, obsessive about labeling guidelines, and always open to opportunities in responsible AI/ML data work.
- LocationIndia
- FocusData · AI/ML
- Graduating2026
- StatusOpen to work
Experience
Data Science & Analysis Intern
- · Performed EDA on real-world datasets with Python (Pandas, NumPy, Matplotlib, Seaborn) — identified patterns and delivered actionable business insights.
- · Cleaned, labeled, and structured multi-format datasets (tabular, text, speech) under strict labeling guidelines and customer-privacy standards.
- · Built interactive Power BI and Tableau dashboards visualizing KPIs and operational metrics for stakeholder reporting.
- · Documented data workflows, proposed process improvements, and maintained accuracy under deadline-driven conditions.
Education
Master of Computer Applications (MCA)
Bachelor of Computer Applications (BCA)
Certifications
- · Google Data Analytics Professional Certificate — Coursera
- · Microsoft Power BI Data Analyst (PL-300)
- · Web Development — SimpliLearn / SkillUP
Projects
Things I've built and shipped.
RetailPulse — AI-Powered Customer Analytics & Demand Forecasting
End-to-end MLOps platform processing 1M+ retail transactions. Prophet + LSTM hybrid ensemble cuts demand-forecast error (MAPE) to 24.1%.
Retailers overstock and understock the wrong items — legacy forecasting misses seasonality, promos, and long-tail SKUs.
Built a full MLOps stack: RFM segmentation (K-Means, DBSCAN), XGBoost churn (0.94 ROC AUC), 5-page Streamlit dashboard, MLflow tracking, Evidently drift detection, Docker/K8s, GitHub Actions CI/CD, Prometheus + Grafana.
More Projects
MediAssist — AI Healthcare Data & Annotation Platform
NLP + LLM-powered assistant that annotates symptom-based medical text across multiple modalities with strict quality control.
Hospital ER Data Analytics — Full-Stack Solution
Automated cleaning & feature engineering on 500+ patient records, advanced MySQL for readmission tracking, and a Power BI dashboard for patient flow and department KPIs.
Real-Time Kafka Streaming Pipeline
Distributed Kafka producer-consumer ETL pipeline processing 10,000+ messages/sec — designed for high-throughput, low-latency multi-format data.
Big Data Analysis — Global E-Commerce Transactions
Analyzed 1M+ transaction records using Apache Spark; 60% faster processing via distributed PySpark aggregation with visualized sales insights.
Always building.
Skills
Let's build something.
Open for AI/ML Data Associate roles, analyst positions, and collaborations. Let's talk.