Ridhiman Sabharwal
Software engineer with experience building scalable backend services and production systems. Skilled in developing reliable APIs, debugging production issues, and improving system performance and reliability in large-scale environments.
Experience
Software Development Engineer Intern
Bright Money
Bangalore, India
–Built and maintained backend services for the USM platform, supporting core user journeys running in production. –Designed REST APIs and asynchronous task workflows in Python using Postgre SQL and Celery, shipping features across user engagement and growth & analytics. –Built an MCP-powered debugging framework integrated with Cursor that automated context gathering, cutting backend debugging time by 60%. –Investigated and resolved production issues, improving reliability of user-facing backend services.
Software Development Intern
Geek Informatic & Technologies
Mohali, India
–Built and deployed REST APIs using Django and Postgre SQL, improving workflow efficiency by20%. –Reduced API latency by30%by optimizing backend query and application logic. –Added OTP-based authentication to secure user-facing endpoints.
Education
Vellore Institute of Technology
B.Tech
CGPA: 8.82
The S.D. Vidya School
Class XII
91.8%
The S.D. Vidya School
Class X
92.6%
Skills
JavaPythonC++JavaScriptSQLHTMLCSSDjangoDjango REST FrameworkReactJWT AuthenticationPostgreMongoDBRedisAWSOracle CloudGitGitHubPostmanLinuxVS CodeMixpanelCeleryTensorFlowscikit-learnNumPyPandasData Structures & Algorithms
Projects
Crypto Predicto
Python, TensorFlow, Pandas
–Built a deep learning model for cryptocurrency price forecasting, reaching85%+ accuracy. –Automated the live OHLCV data ingestion pipeline, cutting processing latency by60%. –Built an interactive dashboard for real-time inference and analysis.
Legal-AI: Bail Prediction
Python, InLegalBERT, Transformers
–Built an NLP system to predict bail outcomes using legal transformer embeddings. –Processed 1,200+ court judgments through an automated ETL pipeline. –Reached a 0.87 weighted F1 score with a hybrid model combining text and structured case metadata.
Yogasana
PoseNet, TensorFlow
–Built a yoga pose classification model reaching95% accuracy. –Ran inference client-side withml5.js, removing the need for server compute. –Worked in an Agile/SCRUM team to ship features iteratively as a finalist among 200+ teams.
Certifications
Azure AI Fundamentals
Microsoft
Data Science Professional
Oracle Cloud
Multi Cloud Professional
Oracle Cloud
Achievements
Chairperson, IE(I)-VIT
Directed chapter operations and grew member engagement by65%.
Coordinator, graVITas’24
Organized121+ eventsfor 20,000+ participants.
Finalist, Women Techies ’24 Hackathon
Top-5 finish out of 200+ teams for building Yogasana.