AI Engineer | Software Developer | Data Scientist
I build Python backends, AI applications and data-driven tools. My work spans APIs, retrieval systems, machine learning pipelines and data analysis using Python, FastAPI, Django and PostgreSQL.
Currently a Data Scientist at Mahindra & Mahindra in Bengaluru. My personal projects explore computer vision, natural language processing, data visualization and web development.
01 / PROFESSIONAL BACKGROUND
Mahindra & Mahindra · Bengaluru
Developing AI applications and backend services, with a focus on retrieval workflows, asynchronous processing and ML platform orchestration.
Mahindra & Mahindra · Bengaluru
Worked on Python backend development, database-backed workflows, retrieval-based recommendations and reporting automation.
02 / SELECTED WORK
COMPUTER VISION / OCR
A Django application for fruit and vegetable freshness detection and product-brand identification, combining image analysis with OCR.
DEEP LEARNING / COMPUTER VISION
A leaf-image classifier with a Django interface. TensorFlow and Keras handle classification, while Gemini generates remedy and report details.
TIME SERIES / DEEP LEARNING
A Streamlit application using an LSTM model to forecast stock trends and visualize predicted values alongside historical prices.
Streamlit sign-in may be required.
NATURAL LANGUAGE PROCESSING
Classifies tweets using TF-IDF and logistic regression. Supports individual text inputs and Excel uploads for bulk sentiment analysis.
Streamlit sign-in may be required.
Selected responsibilities and technologies. Business figures and internal implementation details are omitted.
MULTIMODAL RAG ASSISTANT
Technical reports contain information across text, tables and figures. Finding a relevant answer requires searching lengthy documents, understanding the context and checking the original source.
Built the assistant end to end, covering document ingestion, multimodal retrieval, backend APIs and a conversational interface. Added source citations, conversation memory and processing visibility to help users follow answers back to the supporting documents.
ENTERPRISE WORKFLOWS & REPORTING
Budget preparation involves coordinating reviews across teams, tracking changing estimates and referring to historical information. The application brings these activities together with consistent workflows and reporting.
Served as a lead backend contributor, developing APIs, database-backed workflows and role-based access. Built retrieval-based estimate recommendations and automated presentation reporting, and contributed to review interfaces for the application.
ML PLATFORM & ORCHESTRATION
Physical wheel-force testing requires dedicated vehicle test runs. WFT AI Twin provides a virtual alternative: engineers upload vehicle telemetry CSVs, and machine-learning models predict six wheel-load components (Fx, Fy, Fz, Mx, My, Mz) for each wheel. The platform also supports retraining as new labelled data becomes available.
Developed backend APIs, database access and asynchronous processing workflows. Owned retraining orchestration and contributed to model lifecycle management and results dashboards. My focus was the platform and orchestration; the modeling team developed the predictive models.
03 / TECHNOLOGIES & METHODS
Across professional work, personal projects and data studies.
Python
JavaScript
FastAPI
Django
SQLAlchemy
Pydantic
Celery
Redis
TensorFlow
Keras
scikit-learn
OpenCV
PaddleOCR
NumPy
Pandas
PostgreSQL
MongoDB
BigQuery
PySpark
Apache Airflow
Google Cloud
AWS
Docker
Kubernetes
Helm
Jenkins
Git04 / GET IN TOUCH
For engineering opportunities, collaborations
or a conversation about my work.