About Me

I am a Data Scientist with over 4 years of experience, specializing in AI, Machine Learning, Generative AI, and MLOps. My expertise lies in building scalable machine learning systems that drive real business impact, automating workflows, and optimizing AI solutions for various industries, including Healthcare, Logistics, Sales, and Customer Support.

From predictive modeling and LLM applications to cloud-based AI deployments and robust MLOps pipelines, I thrive at the intersection of technology and business strategy. I've had the privilege of working with industry leaders like Motorola, Caliber Home Loans, and Logitech, ensuring that AI-driven solutions create measurable value.


With a strong focus on agile collaboration, I have led AI/ML teams on various projects, working closely with cross-functional teams to bridge the gap between data science and business objectives. My experience in client interactions has helped me translate complex AI solutions into actionable insights that drive strategic decision-making.


Beyond work, I'm passionate about playing badminton and reading articles on emerging AI trends. I believe in continuous learning and innovation, always exploring new frontiers in AI and automation.

Let's connect and bring AI-driven ideas to life!

Skills

Experience

Education

Certifications

  • Areas of Expertise
    • Machine Learning & AI | Conversational AI & NLP Solutions | MLOps & Model Deployment | Data Modeling & Feature Engineering | GenAI & FinTech Solutions
  • Languages
    • Python, SQL, C/C++, Java, Bash
  • Machine Learning
    • Classification, Regression, t-SNE, PCA, CNN, UNet, OpenCV, NLTK, SpaCy
  • Frameworks
    • Sciki-learn, TensorFlow, PyTorch, Keras, SciPy, XGBoost, Hyperopt, Pandas, NumPy, Vaex, Dask, PySpark, MLflow, Streamlit
  • Data Science
    • Statistical Analysis, Data Mining, Data Wrangling, Data Drift, Concept Drift, A/B Testing, T-Test
  • Generative AI
    • OpenAI, Ollama, Deepseek, Cohere, Mistral, Gemini, Microsoft Phi, Vertax, LangChain, LlamaIndex, HuggingFace, Transformers, RAG (Naive, Rerank, Agentic), LangGraph, NL2SQL, Vector Database (Chroma, Qdrant, Faiss, Pinecone), PEFT, LoRA, DSPy, Explainable AI, Prompt Engineering, Bias Mitigation, Ragas
  • Database Management
    • MySQL, SQL Server, PostgreSQL, MongoDB, Snowflake, Kafka, Salesforce SOQL, Oracle, Elasticsearch, MariaDB, NoSQL
  • Data Visulization
    • Power BI, Matplotlib, Seaborn, Plotly, Dash
  • Cloud & DevOps
    • AWS, Azure, GCP, Sagemaker, Bedrock, Databricks, Hadoop, AWS Lambda, Serverless, Docker, Kubernetes, Airflow, Git, Jenkins, SSH, Github Actions, Infrastructure-as-Code
  • Mar 2022 - Present (Giggso Inc.)
    Data Scientist
  • Dec 2020 - Mar 2022 (Giggso Inc.)
    Junior Data Scientist
  • Aug 2018 - Feb 2021 (Fast NUCES University)
    Teaching Assistant
  • Oct 2019 - Sep 2020 (Fast NUCES University)
    Full Stack Intern
  • National University of Computer & Emerging Sciences (NUCES)
    Bachelor's of Computer Science
  • Government Graduate College
    Intermediate in Computer Science
  • Oracle
    OCI Certified Data Science Professional
  • Databricks
    Advanced Machine Learning Operations
  • Microsoft
    Microsoft Certified: Azure AI Fundamentals
  • Databricks
    Academy Accreditation - Generative AI Fundamentals

My Services

AI & Machine Learning Solutions

Design and deploy intelligent systems for predictive analytics.

Learn more

Generative AI & NLP Services

Create AI-driven content and advanced language models.

Learn more

MLOps & AI Infrastructure

Streamline model deployment and automate workflows.

Learn more

Data Engineering & Business Intelligence

Build robust data pipelines and interactive dashboards.

Learn more

AI Strategy & Consulting

Align AI initiatives with your business goals.

Learn more

Web Scraping & Automation

Extract and process web data at scale.

Learn more

My Work

Generative AI Projects

Mermaid Health

Transforming Healthcare with AI - Simple, Seamless, and Personalized

An AI-powered solution transforming clinical assessment report writing and health management. Mermaid Health automates the generation of comprehensive clinical reports, reducing processing time while enhancing diagnostic accuracy. By leveraging advanced AI models, it streamlines documentation workflows, enabling healthcare professionals to focus on patient care rather than administrative tasks.


Skills: Tensorflow, NLTK, SpaCy, LangChain, LangGraph, Chroma, Docker, OpenAI, Fitz

Factual Inquirer

An AI-Powered Platform for News Credibility & Censorship Assessment

An AI-driven platform that evaluates news credibility and censorship transparency, providing both free and premium insights. Fine-tuned models for News Value (NV) and News Censorship Transparency (NCT) to generate structured JSON outputs with numerical scores and actionable improvement suggestions. Developed a Credibility Named Entity Recognition (CNER) system using OpenAI, leveraging rule-based techniques to rank articles based on credibility parameters. Additionally, built ChatBox, a journalist-focused web application that streamlines article submissions, profile management, peer collaboration, and customer support.


Skills: Transformers, Scikit-Learn, Tensorflow, PyTorch, Keras, HuggingFace, LoRa, PEFT

BidMaster AI

AI-Powered RFP Processing & Proposal Automation

Designed and implemented an AI-driven solution for RFP and RFI processing, automating requirement extraction, classification, and analysis. Developed an intelligent Proposal and Bidding Document Generation system to create structured, high-quality responses, reducing manual effort and ensuring compliance. Leveraging NLP-powered automation, the system enhances accuracy, efficiency, and consistency in the proposal and bidding process.


Skills: NLTK, SpaCy, PyTorch, LangChain, LangGraph, Azure Openai, Fitz, PDFPlumber, Docx

Natural Language Query AI

Conversational AI for Seamless SQL & SOQL Generation

Developed an AI-powered system that enables users to interact with their databases and Salesforce objects using natural language. Implemented query classification to intelligently route user queries to the relevant data sources. Designed and optimized SQL and SOQL generation pipelines using a custom data source schema and RAG models, ensuring efficient and accurate query formation. This solution enhances data accessibility, allowing users to retrieve insights effortlessly without writing complex queries.


Skills: Pandas, LangChain, LangGraph, RAG, Azure Openai, Salesforce APIs

Data Science Projects

Product Demand Forecasting

ML driven time-series product demand forecasting to optimize inventory and business planning.

Developed a machine learning-powered time-series forecasting solution to predict product demand, enabling businesses to optimize inventory management and strategic planning. Implemented advanced ML models to analyze historical sales data, identify trends, and generate accurate demand forecasts. This solution helps businesses reduce stockouts, minimize overstocking, and enhance supply chain efficiency, leading to improved decision-making and cost savings.


Skills: Pandas, NumPy, Sklearn, Matplotlib, XGBoost, HyperOpt

Nuclei Spotter

Detected infected region of Nuclei from divergent images with UNet

Developed a deep learning solution using the UNet model to accurately detect and segment nuclei in diverse microscopic images. Leveraging advanced image processing techniques, the system identifies nuclei across varying conditions, enhancing biomedical research and diagnostic accuracy. This project improves automated cell analysis, supporting applications in medical imaging, pathology, and drug discovery.


Skills: NumPy, Skimage, Matplotlib, TensorFlow

Intent Recognition

Intent Recognition Using BERT on Colab TPU

Developed an Intent Recognition Engine leveraging the BERT (Bidirectional Encoder Representations from Transformers) model to accurately classify user intents from natural language input. Implemented the solution using TensorFlow on Google Colab TPU to accelerate training and inference. BERT's deep contextual understanding pretrained on Wikipedia and large corpora enabled high-performance intent detection across diverse inputs. This project showcases the application of state-of-the-art transformer-based NLP models for real-time language understanding tasks.


Skills: NumPy, Pandas, TensorFlow, Bert, Sklearn, Matplotlib, Seaborn

Recommender System

ML-powered suggestions to resolve application errors

This recommender system is designed to assist in resolving application issues by analyzing error logs and providing context-specific solutions. When an error log is generated, it is first classified as normal or anomalous. If identified as an critical anomaly, the system searches across business documentation, historical group discussions, and pre-configured resolution data stored in Elasticsearch to suggest the most relevant fixes. It returns the top 5 recommended solutions, each with a confidence score, enabling users to take quick and informed action to resolve critical errors efficiently.


Skills: NumPy, Pandas, TensorFlow, BERT, Sklearn, Elasticsearch

Data Engineering & ETL

AnomalyFlow

Scalable Log Pipeline for Smart Anomaly Routing

AnomalyFlow is a scalable data engineering pipeline built to streamline log ingestion and anomaly detection. It ingests data from multiple client servers into Kafka, processes it through an anomaly classification engine, and routes the results accordingly publishing anomalies to hot storage (Kafka and Elasticsearch) for real-time analysis, while directing non-anomalous data to cold storage for long-term retention. Designed to enhance analytical capabilities, AnomalyFlow brings structure, speed, and intelligence to large-scale log processing systems.


Skills: Pandas, SpaCy, NLTK, Sklearn, Kafka, Elasticseach, Kibana

Invoice Processing

Automated Invoice Processing and KPI Reporting

Developed end-to-end ETL pipeline designed to extract and standardize payment receipt data from Citrix ShareFile storage. Each customer folder contains a variety of receipt formats including Images, PDFs, CSVs, and Excel files. The pipeline reads and processes these receipts using OCR for image files and PDF/data parsing tools for other formats. Extracted information is transformed into a standard JSON structure, then loaded into a MySQL database for querying. Users can perform aggregation operations, track business KPIs, and view dynamic reports based on the structured data, enabling smarter financial insights from previously unstructured content.


Skills: Pandas, SpaCy, OCR, PyPDF2, MySQL, Seaborn

Contact Me

mahsan12030@gmail.com

Visit Blog

Copyright © All rights reserved.