May 2024 – Jul 2024 · 2 mos · Kolkata, West Bengal, India · On-site

• ● Enhanced NAS model performance by 15% using Differentiable architecture search (DARTS).
• ● Achieved a 20% increase in efficiency through fine-tuning training steps with custom learning rates and different genotype structures.
• ● Automated environment and Programming setup tasks with shell scripts, increasing workflow productivity by 10%.
• ●Leveraged TensorFlow’s graph-based execution to optimize model computation, improving overall model prediction efficiency by 12% and implement several research papers in the domain of ML(DARTS).

Nov 2023 – Apr 2024 · 5 mos · Remote · Remote

• ● Increased accuracy of electricity consumption predictions by 20% (MSE) across diverse regions.
• ● Handled 2M+ data points efficiently by designing chunk-wise training logic to reduce memory usage.
• ● Reduced model training time by 15% by creating diverse input sequences for model.
• ● Developed LSTM and BiLSTM models, utilizing past 6 years of data to deliver accurate future forecasts, while reducing mean squared error (MSE) by 20%.
• ● Conducted hyperparameter tuning, achieving < 2% absolute error in 75% of the test data and significantly boosting model performance.
• ● Collaborated efficiently with a team of 6 members, leveraging Google Colab and GitHub to develop and manage projects, ensuring seamless integration and effective version control.

May 2023 – Jul 2023 · 2 mos · Remote · Remote

• ● Revamped web app UI, boosting engagement with visually captivating design tweaks, including 10+ GUI changes.
• ● Implemented Frontend form validation, reducing Backend processing overhead by 25% and ensuring data integrity.
• ● Optimized API performance through local storage utilization practices, leading to a 30% reduction in data retrieval latency and improved application responsiveness.
• ● Designed MongoDB Schemas with Mongoose, ensuring 15% more efficient data storage and retrieval.

Apr 2023 – Aug 2023 · 4 mos · Jabalpur, Madhya Pradesh, India · On-site