As an Associate Software Engineer with a B.Tech in Artificial Intelligence and Data Science, I am deeply passionate about building scalable solutions and fostering vibrant tech communities. My technical expertise lies in web development and cloud engineering, with a specialized focus on the Google developer ecosystem—specifically utilizing TensorFlow, Google Cloud Platform (GCP), and Google AI Studio to build intelligent, forward-thinking applications. At Dhruva Advisors India, I engineer AI-driven agentic workflows that automate complex data transformations, dramatically reducing processing time per item from three hours to under twenty minutes. I also redesigned data ingestion pipelines for PostgreSQL, increasing ingestion capacity by over 400%, and spearheaded the migration of a legacy Python application to a modern Nuxt.js framework within a 30-day sprint. These experiences have sharpened my architectural resilience, bridging the gap between heavy backend data processing and seamless web interfaces. My passion for AI also extends to academic research. I have evaluated deep learning architectures to combat adversarial prompt injections on LLMs, achieving a 99.16% detection accuracy using a hybrid CNN and Transformer framework. This work heavily emphasizes Responsible AI (RAI) practices to mitigate algorithmic biases and ensure ethical deployment. Beyond engineering, I am deeply rooted in the Google Developer Groups ecosystem. As the Domain Lead for Testing and Cybersecurity at GDG VNRVJIET, I hosted "HackSprint," a 21-day event introducing ethical hacking to over 120 students. My leadership roles as the Management Coordinator of GDSC VNRVJIET and Founder of VJ Hack/Slash highlight my long-standing commitment to building local ecosystems where developers learn and grow together. I am excited to contribute my technical expertise and community spirit to DevFest, helping create an inspiring experience for developers of all backgrounds.
Associate Software Engineer
Dhruva Advisors india pvt ltd.
- Engineered AI-driven workflows using optimized data structures and LLMs to automate complex tax logic, reducing processing time per item from 3 hours to under 20 minutes.
- Redesigned software architecture and PostgreSQL data ingestion pipelines, increasing backend ingestion capacity by over 400%.
- Led the migration of a legacy Python application to a modern Nuxt.js framework, delivering a fully redesigned frontend within a 30-day sprint.
Indente Web is an individual project developed as a 3-portal React web application (admin, agent, and driver) designed to digitize daily indent ordering and last-mile delivery for wholesale distributors, actively serving over 100 users. Key features and technical accomplishments of the project include: Authentication & Security: Implemented role-based authentication using Supabase for three distinct user roles, featuring portal-aware sessions, phone login for agents and drivers, and middleware route gating. Transport Logistics: Designed a hub-and-spoke transport system with primary and secondary routes, multi-agent stops, drag-and-drop sequencing, and cron-based auto-generating post-indent closure. Ordering Workflow: Built an agent ordering workflow that includes per-company catalogs, cart checkout, and timezone-aware indent closure aligned with the Asia/Kolkata business day. Database Engineering: Engineered a PostgreSQL returnable-container ledger across companies, drivers, and agents. This involved establishing a relational schema for balances, adjustments, and audit trails using multi-table joins, transactional updates, and chunked queries to bypass PostgREST's 1,000-row limit and maintain accurate ledger totals. Analytics & Reporting: Delivered operational reporting comprising nine different views (including HDPOS matrices, trends, and zero-order agents) with PDF and Excel export capabilities. Tech Stack: Built utilizing React, Node.js, Supabase, Next.js, and SQL.
B.Tech in Artificial Intelligence and Data Science
VNR vignana jyothi institute of engineering and technology
Attendee · Oct 24, 2026