Headline

AI Engineer | Chasing rabbit holes

About

Hey, I’m Dhruv Vadhiya. I’m an AI/ML engineer who turns ambiguous ideas into clean, production-grade systems. I’ve built LLM applications, analytics dashboards, speech evaluation tools, and computer vision models — including solutions used by teams for product insights, automation, and real-time decision-making. My work has supported revenue-driving ML products and replaced manual processes with reliable, scalable pipelines. I enjoy taking full ownership of the build cycle: data pipelines, model development, vector search, APIs, and frontend integration. Whether it’s designing a RAG workflow, optimizing models for constrained environments, or spinning up fast prototypes, I focus on clarity, speed, and practical impact. What defines my work: applied ML, LLM engineering, RAG systems, analytics engineering, and the ability to turn rough concepts into polished, working products.

Location

Ahmedabad

AI stack

claude

cursor

elevenlabs

grok

hermes agent

langfuse

Work experience

Software Engineer (G1) at Dexian, Sep 2026 - Present · 1 mo, Bengaluru, Karnataka, India

Artificial Intelligence Researcher at LanguageLoop Technologies Pvt Ltd, Jul 2024 - Oct 2024 · 4 mos, Mumbai, Maharashtra, India, Designed a hybrid retrieval pipeline (BM25 + semantic embeddings) for COBOL analysis, improving top-k retrieval performance from ∼75% to ∼90% across internal benchmarks. Researched and presented LLM + RAG architecture recommendations to founding leadership, covering embeddings, retrieval optimization, and AI-assisted developer workflows.

Summer Intern at Teksun Inc, Jun 2024 - Jul 2024 · 2 mos, Ahmedabad, Gujarat, India, Joined as a trainee and completed a 3-month ML & CNN curriculum in 2 weeks, which led the team to fast-track me onto production-grade work. From there, I designed and built MFCC/STFT-based audio feature extraction workflows on proprietary datasets, then optimized PyTorch CNN inference pipelines for low-latency deployment on edge devices — getting hands-on exposure to how ML systems actually run in production, beyond just model training.

Freelance at QuanticoreAI, Feb 2024 - Oct 2025 · 1 y 8 mo, India, Delivered 7+ AI/full-stack solutions for B2B/B2C clients including NLQ systems, analytics platforms, and AI workflows, generating INR 5L+ revenue. Led a 5-member team, owning AI workflow design, orchestration, fallback strategies, project timelines, code reviews, and client SLAs. Built and deployed custom AI pipelines integrating LLM workflows, REST APIs, and analytics systems with focus on reliability, automation, and production usage.

Education

Bachelor of Technology, Computer Science, SRM Institute of Science and Technology (SRMIST)

Public links

github: https://github.com/dv7453