Skip to content
View hari2594's full-sized avatar
💭
Contributing
💭
Contributing
  • S&P Global
  • Bangalore

Block or report hari2594

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
hari2594/README.md

👋 Hi, I'm Haribaskar Dhanabalan

🚀 Lead Data Scientist | Generative AI & LLM Specialist | AI Strategy & Architecture

I’m a Lead Data Scientist with 9+ years of experience building and scaling enterprise-grade Machine Learning, NLP, and Generative AI systems across EdTech, Retail, Finance, and Mobility domains. I specialize in designing robust AI architectures that move seamlessly from research to production, with a strong focus on reliability, governance, and real-world impact.

My work sits at the intersection of advanced modeling, cloud-native AI systems, and business strategy—helping organizations translate complex problems into scalable, production-ready solutions.


🧠 Core Expertise

  • Generative AI & LLMs: RAG, GraphRAG, LLM fine-tuning, prompt engineering, evaluation & guardrails
  • Natural Language Processing: Text classification, summarization, semantic search, entity extraction
  • Machine Learning & Deep Learning: Supervised & unsupervised learning, model interpretation
  • Knowledge Graphs: Neo4j, AWS Neptune, semantic reasoning & contextual retrieval
  • MLOps & Deployment: Cloud-native pipelines, scalable inference, monitoring & governance
  • AI Leadership: AI strategy, framework design, mentoring, stakeholder communication

🛠️ Tech Stack

Languages Python, SQL, PySpark, R

AI / ML Machine Learning, Deep Learning, NLP, SHAP, LIME, ELI5

Generative AI GPT-3/3.5/4, LLaMA, Mistral, FLAN-T5, BART, Pegasus, LangChain

Databases & Big Data PostgreSQL, Snowflake, SQL Server, Hadoop

Cloud & Platforms AWS (S3, EC2, Lambda, Fargate, SageMaker, Athena)

Visualization & Tools Tableau, KNIME, H2O Driverless AI, RapidMiner, Neo4j


🎓 Education

  • M.Tech – Artificial Intelligence & Machine Learning, IIT Jammu CGPA: 9.5 | Rank 1
  • PGP – Data Science, Praxis Business School, Bangalore
  • B.E – Electronics & Communication Engineering, Anna University, Chennai

🏆 Highlights & Recognition

  • Analytics Vidhya Hackathon – 3rd Place (Insurance Claim Prediction)
  • Happiest Minds – Customer Happiness Award
  • Multiple hackathon wins for applied AI solutions
  • Honorary Associate Chief AI Officer (ACAIO) – Dubai Institute of People

🌱 What I’m Focused On

  • Enterprise adoption of Generative AI & Agentic Systems
  • Reliable, explainable, and governed AI for real-world use cases
  • Scaling LLM systems for accuracy, performance, and trust
  • Mentoring engineers and building strong AI teams

🤝 Let’s Connect


If you’re interested in Generative AI, NLP, or building production-grade ML systems, feel free to explore my repositories or reach out for a discussion.

Pinned Loading

  1. Automated_EDA Automated_EDA Public

    Developed the package to identify categorical and numerical variables of dataset, and plot respective visualization graphs for the variables, identifies missing values and outliers in the variables…

    Jupyter Notebook 6 1

  2. ADAS---Lane-Detection ADAS---Lane-Detection Public

    The problem I have solved in this code is to take a simple video as input data and process it to detect the lane within which the vehicle is moving. Then we will find a representative line for both…

    Python 2

  3. ML-Flow ML-Flow Public

    MLFlow is the generalized package which is developed using the python and the H2o framework. The purpose of this function is to store the models meta data for both classification and regression mod…

    Jupyter Notebook 3 1

  4. Model-Evaluation-Comparison Model-Evaluation-Comparison Public

    This Package is developed to evaluate the classification | Regression models based on different metrics.

    Jupyter Notebook 3 1

  5. PCA---Image-Compression PCA---Image-Compression Public

    The goal of this R Function is to compress arbitrary images using numerical principal component analysis techniques to obtain the most visually appealing compressed image.

    R 1

  6. Vision---EVA Vision---EVA Public