Looking for a job

Dhruv Pandya

@dhruvpandya1

Message

GenAI Engineer at Deloitte (HashedIn) who built RAG automation that reduced manual backend workflows by 80%.

India

Message

At Deloitte (HashedIn), I designed backend services and distributed REST APIs with Django and FastAPI, and worked on GenAI-powered automation using RAG and prompt engineering. That work reduced manual backend workflows by 80% and improved developer productivity.

I also improved service performance by 35% through backend optimization and deployed cloud-native services on Microsoft Azure. Earlier, as a Research Student at Nirma University, I developed an LSTM intrusion detection system and optimized training time by 35% using NVIDIA CUDA.

Experience

Work history, roles, and key accomplishments

Current

GenAI Engineer (Software Development Engineer - 1)

Current

Deloitte (HashedIn)

Jan 2024 - Present (2 years 9 months)

Designed and developed scalable backend services and distributed REST APIs using Django and FastAPI, supporting high-throughput enterprise applications. Engineered GenAI-powered automation and LLM orchestration RAG solutions, improving service performance by 35% and reducing manual backend workflows by 80%.

Django fastAPI Python GenAI LLM RAG Prompt Engineering Claude OpenAI Azure Agile Root Cause Analysis REST APIs Testing Monitoring Access Control

Software Development Engineer Intern

Deloitte (HashedIn)

Dec 2023 - Dec 2023 (0 months)

Built backend-driven applications using Spring Boot and PostgreSQL, delivering complete systems such as a Leave Management System and Restaurant Booking System. Designed and optimized RESTful backend APIs and implemented data processing pipelines using PySpark and Apache Airflow.

Spring Boot PostgreSQL REST APIs PySpark Airflow Docker CI CD Backend Development

Research Student

Nirma University

Jan 2023 - Jun 2023 (5 months)

Developed an intrusion detection system using LSTM models on the KDDCup dataset, improving threat detection accuracy through deep learning techniques. Optimized model training pipelines with NVIDIA CUDA, achieving a 35% reduction in training time.