Hi there! 👋 I'm Ashhar Farooqui, a passionate Data Scientist & ML Engineer based in Bengaluru, India. My journey is one of curiosity-driven discovery where I transform complex datasets into compelling narratives that drive business decisions and solve real-world problems.

"Data is the new oil, but insights are the engine."

My career has been defined by one core mission: extracting meaning from chaos. Whether it's predicting market trends, uncovering patterns in time series data, or building intelligent systems that learn and adapt, I believe that every dataset has a story to tell. My role is to be that storyteller—armed with Python, machine learning, and a relentless curiosity.

• 🧠 Solving Complex Problems with machine learning and statistical rigor
• 📊 Telling Data Stories that influence strategic decisions
• 🚀 Building Scalable Solutions that create real-world impact
• 🔍 Continuous Learning in the ever-evolving AI/ML landscape
• 🤝 Collaborating with Teams to deliver data-driven products

Domain	Technologies
Programming
ML/DL
Data Stack
Visualization
Web/Apps
Cloud/Infra
Analytics

Forecasting crude oil prices using machine learning and time series models

The Challenge: Can we predict volatile commodity prices with acceptable accuracy?

The Solution:

• 🤖 Built an ensemble of 5 sophisticated models (ARIMA, SARIMA, Prophet, LSTM, XGBoost)
• 📊 Achieved 97% R² score with ±$2-3/barrel prediction accuracy
• 🎨 Created an interactive Streamlit dashboard for real-time insights
• ⚙️ Implemented automated retraining pipelines for model freshness

Key Metrics:

• RMSE: 1.65 | MAE: 1.23 | MAPE: 1.72%
• Successfully captures 89% of price direction changes
• Backtested on 12 months with 93% out-of-sample accuracy

Tech Stack: Python, TensorFlow/Keras, XGBoost, ARIMA, Prophet, Streamlit, Plotly

🔗 View Repository | Read Full Documentation

• 🤖 Supervised Learning: Regression, Classification, Ensemble Methods
• 🔄 Unsupervised Learning: Clustering, Dimensionality Reduction, Anomaly Detection
• ⏰ Time Series Analysis: ARIMA, SARIMA, Exponential Smoothing, Prophet, LSTM
• 🧠 Deep Learning: Neural Networks, CNNs, RNNs, LSTMs, Transformers

• 📊 ETL Pipelines: Design, development, and optimization
• 🗄️ Big Data Processing: Spark, Hadoop, distributed computing
• 📈 Statistical Analysis: Hypothesis testing, A/B testing, Bayesian methods
• 🔍 Data Quality: Validation, profiling, and governance

• 📝 Text Classification: Sentiment analysis, toxicity detection, topic modeling
• 🗣️ Language-specific NLP: Multilingual support (English, Hindi, Bengali)
• 💬 Sequence-to-sequence Models: Translation, summarization
• 🔎 Information Extraction: Named entity recognition, relation extraction

• 📊 Dashboard Development: Tableau, Power BI, Streamlit
• 📈 KPI Monitoring: Real-time analytics and alerting
• 💡 Data Storytelling: Translating insights into business narratives
• 🎯 Predictive Analytics: Forecasting and trend analysis

I'm passionate about continuous learning and staying at the forefront of data science innovation:

• 📖 Regular reader of arxiv papers on ML/AI
• 🎓 Coursework in Advanced Statistics, Deep Learning, Time Series Analysis
• 🔗 Active contributor to open-source projects
• 🎤 Enthusiast in technical blogging and knowledge sharing
• 🤝 Mentor in ML communities helping aspiring data scientists

I'm always excited to discuss:

• 🤔 Novel approaches to complex data problems
• 🚀 Building scalable ML systems in production
• 📊 Data storytelling and visualization techniques
• 🔬 Latest research in AI and machine learning
• 🌍 Real-world applications of data science

I'm currently exploring:

• 🤖 Large Language Models and their applications
• 🔮 Generative AI for data synthesis and augmentation
• 🌍 Federated Learning for privacy-preserving ML
• 📊 Causal Inference for understanding root causes
• ⚡ Edge AI for real-time, on-device predictions
• 🔗 Graph Neural Networks for relational data

I believe in giving back to the community:

• 🔧 Active contributor to scikit-learn, TensorFlow, and Pandas projects
• 📚 Maintainer of several ML utility libraries
• 💡 Creator of data science templates and boilerplates
• 🎓 Author of tutorials on advanced ML topics

┌─────────────────────────────────────────────────────┐
│  Data  →  Questions  →  Exploration  →  Modeling    │
│                                             ↓       │
│  Validation ← Interpretation ← Results ← Training   │
│       ↓                                             │
│  Insights → Storytelling → Action → Impact          │
└─────────────────────────────────────────────────────┘

1. Curiosity First 🔍 — Always ask "why" before jumping to solutions
2. Data Integrity 🛡️ — Garbage in, garbage out; prioritize quality
3. Interpretability 🧩 — Black boxes are the enemy; explain your models
4. Reproducibility 🔄 — Make your work auditable and repeatable
5. Impact-Driven 🎯 — Every model should solve a real problem
6. Continuous Iteration 🚀 — V1 is never the final version

• 📘 ML Best Practices Guide — Industry-standard approaches
• 🔬 Time Series Playbook — Comprehensive forecasting guide
• 📊 Data Visualization Cookbook — Creating impactful visuals
• 🧠 NLP Starter Kit — Text analysis from scratch
• 🚀 Production ML Checklist — Deploying models reliably

• 🧠 Specialization: Time Series Analysis & Forecasting
• 💼 Experience: 5+ years in Data Science & ML Engineering
• 📍 Location: Bengaluru, India 🇮🇳
• 🎯 Goal: Building intelligent systems that solve real problems
• 🌱 Growing Interest: Generative AI, MLOps, Causal Inference
• ☕ Fun Fact: I debug code with coffee and classical music! 🎵

If you're interested in:

• 🤝 Collaboration on data science projects
• 💼 Discussing job opportunities
• 🎓 Learning from my experience
• 🔗 Building something amazing together

Feel free to reach out! I'm always up for interesting conversations and exciting challenges.