Siddhant Patil

Experience

• 

  CSIR-CEERI

  May 2017 - Jul 2017

  Summer Intern

  • Group Leader• Area of Work: Image Processing and Machine Learning• Designed and developed an application suite for a fully automated fabric analyzer which sorts fabric based on color after accounting for the change in color temperature using Python, C# and Microsoft Visual Studio. (IEEE - Conference Paper)• Won Mantra Award - Innovator of the Year 2017-18

• 

  Birla Institute of Technology and Science, Pilani

  Aug 2017 - May 2018

  Course: CS/EEE/INSTR F241 Microprocessors and Interfacing Course: Introduction to Cadence Virtuoso Areas of Work: 1.) Designing Voltage Amplifiers - IEEE Conference Paper 2.) Design, Implementation and Application of Vedic Multipliers - Journal Paper Course: BITS F112 Technical Report Writing

  • Teaching Assistant

    Jan 2018 - May 2018

  • Course Instructor

    Jan 2018 - May 2018

  • Undergraduate Student Researcher

    Aug 2017 - May 2018

  • Teaching Assistant

    Aug 2017 - Dec 2017
• 

  IEEE - International Conference on Devices, Circuits and Systems

  Mar 2018 - Mar 2018

  Chairperson

  • 1 selected from 1000+ applicants based on research experience• Guided 16 researchers by finding the missing gaps in their research work and showing them the future path

• 

  Technische Universität Dresden

  May 2018 - Dec 2018

  Software Engineer Intern

  • Department: Informatik (Computer Science)• Area of work: Approximate Computing - Design Space Exploration of Approximate Digital Signal Processing Systems (applicable to Machine Learning algorithms)• Fully funded by the Chair for Processor Design

• 

  Birla Institute of Technology and Science, Pilani

  Jan 2019 - May 2019

  Teaching Assistant

  Course: EEE/INSTR F341 Analog Electronics

• 

  Hiroshima University

  Jun 2019 - Jun 2019

  Intern

  • ILDP-START plus Program provides an introduction to the exciting and emerging field of cutting-edge technology of Hiroshima University, especially the laboratories relevant to ILDP target fields.• Fully funded by ILDP START plus 2019 Program and Sakura Science Plan, Japan

• 

  Quantiphi

  Jul 2019 - Aug 2021

  • Document Classification:Trained a document classifier having 99.99% accuracy on the client dataset using the Random Forest Classifier.• Document Digitization: -- Trained object detection models using ssd-inception-v2-coco, ssdlite-mobilenet-v2-coco and faster-rcnn-resnet50-coco. Improved the performance by increasing the dataset size and variety using customized data crawling and data augmentation. -- Built a pipeline for digitizing images of tables, graphs, plots, P&I diagrams that gave .json/.csv/.docx as output. -- Converted a scanned pdf into text select-able pdf. -- Performed pre-processing on scanned documents for correcting orientation using image processing algorithms. -- Managed the tagging team using the Google Data Labeling Tool.• Analytics on the digitized document: -- Entity extraction: Trained custom models for extracting key-value pairs with 96%+ accuracy. As an alternative utilized Google Cloud Products, namely, Google Document Understanding AI (DUAI) - Invoice Parser, Form Parser and Table Parser in raw information extraction and applied post-processing rules to increase the accuracy of key-value pair extraction from 81.6% to 94.2% on client invoices. -- Relevant passage retrieval: Extracted relevant passages from a document using the user queries and a custom pipeline that used one of GloVe, DistilBERT and BERT embeddings with Cosine Similarity. Also, used Manhattan distance as a similarity measure on the LSTM model. -- Text Summarization: Developed a solution to provide a concise summary of the document using robust techniques. -- Topic Modeling: Used Latent Dirichlet Allocation for extracting relevant topics from a digitized document.• Deployment: -- Implemented parallel processing to reduce the latency of the pipelines. -- Performed training and testing on the Google Compute Engine and deployed the pipelines on Google Cloud Function, Google App Engine and Google Cloud Run. Show less

  • Senior Machine Learning Engineer

    Mar 2021 - Aug 2021

  • Machine Learning Engineer

    Jul 2019 - Feb 2021
• 

  University of California, Los Angeles

  Sept 2021 - now

  Course: Introduction to Laboratory and Scientific Methodology Course: Introduction to Laboratory and Scientific Methodology Course: Introduction to Laboratory and Scientific Methodology

  • International Student Ambassador

    Jun 2022 - now

  • Graduate Teaching Assistant

    Mar 2022 - Jun 2022

  • Graduate Teaching Assistant

    Jan 2022 - Mar 2022

  • Graduate Teaching Assistant

    Sept 2021 - Dec 2021
• 

  Amazon

  Jun 2022 - Sept 2022

  Data Scientist Intern

  • Designed a knowledge-based recommender system by ensembles topic modeling and cosine-based similarity on sentence-level sentence embedding to improve NDCG@5 by 13.8%.• In topic modeling replaced HDBSCAN with GMM for exhaustive clustering to improve Silhouette score by 11.7%.

• 

  Synopsys Inc

  Sept 2022 - Dec 2022

  Machine Learning Software Engineer Intern

  • Improved the performance / visualization of the Synopsys Design Space Optimization (DSO) Tool using Django and ML.

• 

  Cadence Design Systems

  Jan 2023 - now

  • Software Engineer II

    Apr 2023 - now

  • Software Engineer Intern

    Jan 2023 - Mar 2023