Open to opportunities

Dhruv Sood

@dhruvsood

Message

ML-focused software engineer building production-grade RAG systems and high-performance retrieval infrastructure.

India

Message

What I'm looking for

I seek roles where I can build production ML systems (RAG, search, ETL) with strong engineering practices, measurable impact, and collaborative research-driven teams.

I am an ML-focused software engineer who builds production-grade retrieval-augmented systems, scalable ETL pipelines, and high-performance search infrastructure. I have architected RAG multi-agent chatbots with real-time streaming, engineered OpenSearch-backed retrieval layers that improved latency by 96%, and scaled content platforms to thousands of daily users.

My work spans LLM fairness research, vector databases, and enterprise ML deployments — from designing Qdrant/LlamaIndex retrieval systems to automating Airflow pipelines processing 100K+ records daily. I combine strong software engineering, ML research experience, and system design to deliver measurable product impact and robust, production-ready solutions.

Experience

Work history, roles, and key accomplishments

Current

AI Research Intern

Current

Artificial Intelligence Institute of South Carolina

Feb 2025 - Present (8 months)

Collaborating on LLM fairness research to mitigate caste and religious bias; developed bias detection algorithms and curated a 46K-prompt AI Constitution dataset to benchmark GPT-4, LLaMA, and Indic models.

Python PyTorch Benchmarking LLaMA GPT 4

Current

Software Engineer

Current

Probo

Jan 2025 - Present (9 months)

Architected a production-grade RAG multi-agent chatbot with SSE and entity extraction, boosting user engagement and page visits by 35%, and optimized retrieval infrastructure to cut query latency from 1.5s to 60ms (96% improvement).

RAG OpenSearch SQL Python Apache Airflow ETL Vector Databases Real Time Widgets

ML Engineering Intern

National Stock Exchange (NSE India)

May 2024 - Jul 2024 (2 months)

Designed an enterprise RAG system using Qdrant and LlamaIndex that improved regulatory document retrieval accuracy by 50% and ingested 10K+ documents with hybrid semantic and keyword search.

Qdrant LlamaIndex Vector Search Python Semantic Search Information Retrieval

Education

Degrees, certifications, and relevant coursework

Indraprastha Institute of Information Technology, Delhi

Bachelor of Technology, Computer Science

2021 - 2025

Grade: 8.03/10

Activities and societies: Gold medalist, state topper in NGSE, high ranks in JEE Mains, CodeChef, Codeforces; mentor and NGO volunteer; participated in competitions such as Adobe GenSolve and IEEE BlackSlash.

Bachelor's in Computer Science with specialization in Social Sciences (Finance/Economics); completed coursework in data structures, machine learning, NLP, information retrieval, and databases with a CGPA of 8.03/10.