Headline

Turning complex numbers into smarter real-world decisions | AI & Data Science Undergraduate @ LSPGCOE | Passionate about accessible technology.

About

Numbers fascinate me—especially when they help shape smarter decisions. I enjoy turning data into clear, practical insights that make real-world impact. I’m Abhishek Bhosale, a final-year AI & Data Science student and self-taught coder with a 5★ Python badge. I build end-to-end ML solutions focused on usability and impact. My projects include a Credit Risk Model using SMOTE-Tomek sampling with live predictions via Streamlit, a Health Insurance Premium Estimator that predicts fair pricing from lifestyle data, and PASE, where I built two Random Forest models to balance wind energy supply with city demand, deployed through FastAPI for real-time use. I work confidently with Python, Pandas, NumPy, scikit-learn, XGBoost, along with Git, VS Code, Google Colab, and Streamlit deployments. I’m certified by HackerRank, Lloyds Banking Group, AWS, and the Goldman Sachs Virtual Experience Program. Beyond code, I value clear communication—sharpened through quiz and debate wins at GDG events. Driven by curiosity and clean code, my goal is simple: make technology accessible and insights actionable across finance, healthcare, and energy. Let’s connect if you’re building something where data can make a difference. Open to Junior Data Scientist or ML Engineer roles"

Location

Tasgaon

AI stack

cursor

langchain

langgraph

openai

Work experience

Intern at Innomatics Research Labs, Feb 2025 - Present · 1 yr 5 mos, Engineered text-cleaning pipelines (deduplication, normalization) to format unstructured datasets for Large Language Model (LLM) ingestion, improving input data quality. Wrapped pre-trained Generative AI models into RESTful APIs using FastAPI, enabling structured batch inference and scalable deployment.

Project Trainee at KONKAN RAILWAY CORPORATION LIMITED, Feb 2025 - May 2026 · 1 yr 4 mos, Ratnagiri, Maharashtra, India, Analyzed existing railway database schemas, identifying table relationships and primary/foreign keys to map out the underlying logic for an automated query generation tool. Built a No-Code SQL Generation Engine [GitHub Repository] using Python and PyPika that reads database metadata to automatically generate complex queries (joins, filters, aggregations) for non-technical staff. Developed FastAPI endpoints that process user parameters and return formatted SQL syntax, significantly reducing manual SQL query writing for non-technical staff.

Machine Learning Intern at DataPointer Technologies, Sep 2025 - Oct 2025 · 2 mos, Processed 10,000+ real-world insurance records, applied EDA to eliminate approx. 30% redundant features, directly reducing model inference time and improving input data quality. Trained and benchmarked Linear, Ridge, and XGBoost regressors via cross-validation for insurance premium prediction, selecting the optimal model based on production-grade performance metrics.

Education

Bachelor of Technology, Artificial intelligence and data science, Dr. Babasaheb Ambedkar Technological University