Aditya Dhookia Email & Phone Number

Bengaluru, Karnataka, India

Pune, Maharashtra, India

Fast Track - 6 months ahead of track.Building Speech and LLM’s Agents for Financial Databases, Confluence.Built ML models for liquidity forecasting and management.

• Working on some cool projects. • Developed smart contracts on blockchain tech.• Developed Big data pipelines on Spark for critical feed enrichments for organization profit and loss data. • Awarded a bank-wide performance excellence award - "Recognition Award 2022" - for exceptional performance. • Mentored 2 interns (10 weeks) to deliver a smart recruiter tool that aims end-to-end Resume Parsing and Finding Candidates for a Job Description using BERT. • Developed a platform that can convert JSON/Avro format data to CSV/excel up to 20 lakhs rows of data, with the capability to perform SQL queries and integrated it with Hadoop storage.

Paris, Ile-De-France, France

1. Prepared a new dataset from scratch containing more than 5000 labeled images which include depth ordering information.2. Designed a CNN based Encoder-Decoder architecture for accurate depth estimation for occluding objects from a single image.3. Worked on Development of a Efficient Labelling tool for adding depth ordering information especially for Occluding Objects.

Switzerland

Worked on Investigating the potential of weakly supervised segmentation within the context of CT lung image analysis aiming at detection of interstitial lung disease (ILD) related patterns.

1. Developed a new algorithm for Early Stage Dementia Detection based on Recurrent Neural Networks (RNN)(LSTM and GRU) and achieved state-of-art performance on ARUBA dataset.2. Worked on Application of Unsupervised Domain Adaptation by Backpropagation to overcome the insufficiency of Neuroimaging Data.

South Korea

1. Implemented an efficient algorithm for real time low resolution face recognition based on Discriminant Correlation Analysis (DCA).2. Worked on solving airlight estimation problem using internal recurrence property of images for the task of image dehazing.