Indrajeet Aditya Roy

Experienced Software Engineer skilled in designing APIs, data pipelines, and machine learning solutions in cloud environments. Proficient in Python, TensorFlow, and PyTorch. Strong background in backend development and system optimization. Demonstrated ability to improve scalability and reliability in fast-paced technology environments.

Resource-Aware Task Scheduling for Mobile Edge Cloud Computing
– Developed and benchmarked a distributed scheduling system for mobile edge cloud task workload orchestration, implementing a dynamic resource
aware task allocator in Python using Reinforcement Learning (Q-Learning), and optimization algorithms (HEFT).
– Achieved 20% reduction in mobile device energy consumption and improved task processing schedule deadline compliance on simulated mobile workloads.

Natural Language to SQL for Database Exploratory Data Analysis
– Fine-tuned OpenCodeInterpreter-6.7B LLM using LoRA+ and QLoRA (PEFT) on 78K examples for Text-to-SQL generation, optimizing training to 4 hours on single A100 GPU with PyTorch and Hugging Face Transformers.
– Built production pipeline integrating with Databricks/Snowflake data warehouses, achieving 85%+ query accuracy and reducing SQL query development time.

XML Data Warehouse for Data Forecasting and Analysis
– Built data warehouse using Python, PostgreSQL, and Apache Spark with hybrid OLTP/OLAP architecture, implementing parallel XML processing via SAX parsers, multithreading, and snowflake schema design for data analytics.
– Automated data loading and extraction pipelines, achieving a 68% reduction in ETL time and enabling sub-second OLTP queries on 30GB datasets.