Vedansh Dubey

About Vedansh Dubey

I build production-grade AI systems that operate under real-world scale and constraints.As a Founding Member at Melooha, I work across the full AI product stack — from model integration and NLP system design to backend architecture and frontend application-layer deployment. My focus is on building ML-powered features that are not only accurate, but also performant, scalable, and cost-efficient.My work involves:• Designing and optimizing AI inference pipelines• Integrating NLP and transformer-based models into production workflows• Engineering scalable backend systems and APIs to support concurrent usage• Improving model performance through iterative experimentation and evaluation• Reducing inference latency and infrastructure cost through architectural optimizations• Contributing to cross-platform application systems that deliver AI-driven experiences reliablyI am particularly interested in the intersection of applied ML and systems engineering — where model quality, infrastructure design, and user-facing performance must align.Previously, during my AI/ML research internship, I worked on NLP and transformer optimization for sentiment analysis in Indic languages, focusing on improving model robustness and efficiency.My core interests include:• Scalable AI system design• Inference optimization & performance engineering• End-to-end AI product development• Applied NLP & Production grade MLI am driven by building AI systems that solves a real world problem, creates impact in user life & function reliably in production — not just in experimentation environments.

Experience

1. Founding Member

   Melooha

   Jun 2024 — Present

   Leading the architecture and development of a production-grade AI-driven astrology platform serving 2M+ users closely with the CTOScaling the system from early-stage MVP to high-traffic infrastructure, addressing performance, reliability, and concurrency challengesTook ownership of multiple high-impact (“hero”) features — driving backend implementation, AI integration, testing, deployment, and post-release optimizationDesigned and optimized AI inference pipelines to support large-scale concurrent user requests with low latencyImproved model accuracy through iterative experimentation, prompt engineering, data refinement, and transformer-based optimizationsBuilt and enhanced scalable backend systems (Python), REST APIs, and database architecture in coordination with the engineering teamContributed to cross-platform application development (Flutter), ensuring smooth mobile performance at scaleImplemented caching mechanisms, asynchronous processing, and compute-efficient execution strategies to reduce inference latency and infrastructure costHelped establish monitoring, logging, and analytics frameworks to maintain production stability and track model performanceContinuously shipped AI-driven feature improvements while maintaining system reliability

Education

• Dinanath High school and Jr. College, Nagpur

  HSC, PCM

• Indian Institute of Technology, Ropar

  Majors, Artificial Intelligence

• Sunny's Spring Dale School, Bhandara

  SSC

• Shri Ramdeobaba College of Engineering and Management

  Bachelor of Technology - BTech, Information Technology