AI/ML Engineer | Data Scientist
Where Data, Machine Learning, and AI meet Software Engineering.
From data science and predictive modeling to modern AI applications and scalable backend systems, I build reliable solutions designed for real-world impact.
Currently exploring Agentic AI · Multimodal Systems · AI Infrastructure
I'm Sweety, an AI/ML Engineer with a strong foundation in Statistics, Mathematics, and Computer Science. I enjoy building software that combines data, machine learning, and thoughtful engineering to solve real-world problems.
My work spans data science, machine learning, backend engineering, and AI applications. From data preparation and model development to APIs, deployment, and production systems, I enjoy working across the entire development lifecycle while keeping solutions practical, scalable, and maintainable.
I focus on problem-solving, clean architecture, and engineering best practices rather than chasing trends. Outside of work, I build personal projects, contribute to open source, and write technical articles to deepen my understanding and share what I learn.
I build intelligent, data-driven applications — from end-to-end ML pipelines to agentic AI systems. Here are some of my projects.
AI-powered financial document intelligence platform built with Retrieval-Augmented Generation (RAG). EquityLens helps equity research analysts query, compare, and analyze Annual Reports, Quarterly Results, Investor Presentations, and Earnings Call Transcripts using natural language, delivering grounded answers with page-level citations.
A multi-tenant, agentic platform that aggregates, filters, and curates tech news and video transcripts from YouTube, blog and RSS feeds. The system scores and ranks incoming articles against personalized user profiles, delivering tailored daily digests directly to users' inboxes via an automated delivery pipeline.
Built a real-time cryptocurrency flash crash prediction platform using Online Machine Learning that streams live market data, engineers statistical features, continuously updates an Adaptive Random Forest model with ADWIN concept drift detection, and exposes predictions through a FastAPI REST API.
Deep Learning application using U-Net architecture for automatic semantic segmentation of lung nodules from CT scan images. Features an interactive Flask web UI that processes input scans to generate binary segmentation masks and color-mapped jet overlays for clinical visualization.
Production-grade ML app with a dual-model architecture: one classifier for loan approval, one regressor for sanction amount. Built with a 5-stage modular pipeline (ingestion → validation → transformation → training → evaluation), Flask REST API, and deployed on Render with Docker.
Building AI systems in production environments, end-to-end.
NovarsisTech Pvt. Ltd.
Strong foundations in Mathematics, Statistics, and Computer Science — the backbone of my journey into AI and Machine Learning.
Master of Science - MSc., Computer Science
Completed my Master's degree in Computer Science at NIT Trichy with a CGPA of 7.62. My coursework covered Data Structures, Algorithms, Object-Oriented Programming, Database Management Systems, Operating Systems, Computer Networks, and advanced topics in AI, Machine Learning, and Data Science.
Bachelor of Science - BSc.- Mathematics & Statistics
Completed my Bachelor's degree in Mathematics, Statistics and Economics at Govt. Holkar Science college Indore with 81.14%. My coursework covered Statistical Techniques, Probability Theory, Hypothesis Testing, Time Series Analysis etc.
XII, PCM
I completed my class 12 high education from MPBSE, where I studied Science with Mathematics, Physics and Chemistry.
Core competency areas built through real production work, not tutorials.
The tools I reach for when building production AI systems.
LangGraph isn't just another agent framework — it's a paradigm shift. Here's how state machines make your AI agents reliable and debuggable.
Retrieval-Augmented Generation sounds simple. But production RAG has 12+ failure points. Here's every one of them explained.
A collection of Python concepts that even experienced developers often overlook — generators, context managers, dataclasses and more.
I'm always open to new opportunities and collaborations. Feel free to reach out directly at sweetymahale23@gmail.com
