Harbhajan Singh
I’m a data scientist who loves turning complex data into clear, actionable insights and building scalable analytics solutions.
About
I’m a data scientist who loves turning complex data into clear, actionable insights and building scalable analytics solutions. My work spans finance, government, and tech, where I’ve delivered end‑to‑end data pipelines, forecasting models, and AI-powered decision support. I enjoy collaborating with cross‑functional teams to drive business impact and continuously learn new methods.
Currently, I contribute at TD Securities, building real‑time analytics and ML-powered insights, while also engaging in RAG/LLM initiatives and OCR/text extraction projects. I’ve led diverse initiatives from data ingestion to dashboards and predictive analytics, with experience across on‑prem and cloud environments.
- 7 years' experience
- Joined February 2026
Skills· 15
Languages
- EnglishAdvanced
Industry experience
- Financial Services
- Government
- Software & Internet
- Professional Services
Work experience· 8 roles
Built a PowerApps-based real-time pipeline monitoring application, delivering $700K in overall cost savings and improved visibility into deal activities. Designed end-to-end data pipelines for ingestion, transformation, and validation to support scalable analytics. Created interactive dashboards and KPI visualizations for real-time business insights. Developed forecasting models, predictive analytics, and sentiment analysis solutions to inform decision‑making across origination teams.
Contributed to a RAG-based AgPal chatbot, automating test scripts with Azure OpenAI services (Prompt Flow, Semantic Kernel, LangChain). Evaluated LLM responses against GroundTruth and Groundedness scores. Implemented deduplication tool improvements, and led data collection for arthropod detection. Created a dashboard to monitor project status and boosted team productivity by 30%.
Led end-to-end development of a Retrieval-Augmented Generation (RAG) product using OCR to extract text from PDFs (including scanned documents). Applied BERT-based data correction to fix spelling errors and stored results in a database. Built a Streamlit UI integrating LLMs (e.g., GPT-3.5-turbo) to answer user queries based on a connected knowledge base.
Education
Certifications and awards
Hire Harbhajan Singh today
Post a job and invite Harbhajan, or contact them directly. No hidden commission.







