I'm Thomas

Data Analyst

FRVR

• Optimised SQL queries to enhance efficiency, reducing BigQuery costs by 40%.
• Designed and developed interactive dashboards in Sigma and Looker to provide key insights for stakeholders.
• Built scalable dbt data models by integrating MongoDB, Google Sheets, and BigQuery to streamline pipelines and boost analytics performance.
• Automated game data extraction with Selenium and Apache Airflow, eliminating manual collection and improving reliability.
• Developed reusable BigQuery UDFs by integrating Python, Pandas, NumPy, GCP Cloud Run, Vertex AI, FastAPI, and Gemini to enable scalable data processing.
• Built a chatbot using the MCP framework, Next.js, FastAPI, LangChain, Gen AI and Google AI Studio to enable natural language queries, allowing users to generate charts and retrieve data seamlessly.

Software Engineer

EasyHiring

• Built and optimised a scalable Learning Management System (LMS) with Next.js, Tailwind CSS, and FastAPI, improving performance and responsiveness.
• Integrated a distributed queue system with Celery and RabbitMQ, enabling reliable background processing and real-time notifications, which reduced system latency and improved user experience.
• Implemented secure user authentication and authorisation with Clerk, strengthening data protection and access control for users.
• Supported deployment and maintenance of microservices on GCP, helping streamline release cycles and increase overall system stability and uptime.

Software Engineer

Freelance

• Assisted an analyst with the pioneering process automation within the Investments Department for STIF entries, Position recon, Trade Costs, & Matching.
• Developed a comprehensive pipeline using Microsoft Graph API, Microsoft's Power Automate, Microsoft Azure Functions, Python, Selenium, and FAST API to automate various tasks. These included searching for files within emails and web portals, extracting information from PDFs and Excel documents, converting data into JSON and XML formats, performing additional calculations, and storing this data in a database and Excel files for future use.
• This significantly reduced the workload for analysts and associates by automating tasks that were previously carried out manually, resulting in substantial time savings.

Software Engineer

Ynoox

• Successfully designed and published an Automation tool for QNAP's NAS using Python, Django Framework, SQL, JavaScript, Linux Shell Scripting, RabbitMQ, and Docker. Thus, automating the task saves the user time. The application focused on editing and extracting data from PDFs. The app was packaged and published on QNAP’s App Store.
• Worked on a document classifier model to classify PDF documents and a custom entity extraction model to extract information from document texts using BERT, Python, NLP, PyTorch, Pandas, Numpy and Spacy.
• Collaborated on building Node JS, React, Webpack, JavaScript, TypeScript, Kafka, Flask, Git and Python-based web applications as per the client's requirements.
• Maintaining the PF4me's Python Libraries.
• Created Pdf4me automation modules for Integromat (Make) and Zapier.
• Providing support to customers by resolving any bugs and adding new features to the application.

University Project

Airport Authority of India (Nagpur)

• Collaborated to Create a centralized log system for ATCO officers of Airport authority India.
• Built the backend for the website using PHP and MySQL.
• Managed the Team by assigning the task to team members and