A Jadhav

AI Researcher

Bengaluru, Karnataka, India3 yrs 9 mos experience

Most Likely To SwitchAI Enabled

Key Highlights

1.5 years of hands-on experience in AI and Data Science.
Awarded Best Paper at Philips Global Data & AI Conference.
Expertise in Computer Vision and Natural Language Processing.

Stackforce AI infers this person is a Data Science and AI professional with a focus on Healthcare applications.

Contact

Skills

Core Skills

Natural Language Processing (nlp)Computer VisionMachine LearningData Engineering

Other Skills

Advanced MathematicsArtificial Intelligence (AI)BERTCluster AnalysisComputer ScienceData AnalyticsData CleaningData PipelinesData PreparationData ProcessingData ScienceData StructuresData VisualizationDatabasesDatasets

About

Hello World! This is Ashwath. With over 1.5 years of hands-on experience in AI and Data Science, I currently serve as an AI Engineer at Qualcomm. During my tenure here, I have successfully spearheaded three impactful projects. One of these projects delved into the realm of Computer Vision, focusing on Optical Character Recognition (OCR), while the remaining two showcased my proficiency in Natural Language Processing (NLP). Prior to my role at Philips, I contributed as a Machine Learning Intern at AMD, where I completed several ML projects, honing my skills and gaining industry insights. Master's degree in AI/ML from the prestigious Indian Institute of Information Technology, Lucknow.

Experience

Qualcomm

AI Software Engineer

Mar 2024 – Present · 2 yrs · Hyderabad, Telangana, India · Hybrid

Philips

AI Engineer

Jan 2023 – Mar 2024 · 1 yr 2 mos · Bengaluru, Karnataka, India · Hybrid

De‑Identification of PHI in Client Narrative: Achieved 92% accuracy in de-identifying PHI in unstructured text data by combining
Clinical BERT and BIO BERT models and fine-tuning them to perform NER.
Cross‑lingual Transferability of Monolingual NER Model: Designed a model capable of accurately identifying PHI in Spanish and 3 more
languages, despite the absence of annotated NER data for the Spanish language.
Obtained an accuracy of 84% by fine-tuning the embedding layer of the BERT with a Spanish corpus and the other layers with an NER dataset.
Earned the prestigious Best Paper Award in the NLP category at the 2023 Philips Global Data & AI Conference for the project.
OCR Model: Explored OCR models, including Easy-OCR, Paddle-OCR, Keras OCR, and Tesseract, to identify the most suitable for the project.
Created synthetic data to expedite training, addressing the time-consuming task of labeling real data, Selected and fine-tuned Easy

Python (Programming Language)PyTorchBERTNatural Language Processing (NLP)Computer Vision

Amd

Machine Learning Intern

Jul 2022 – Jan 2023 · 6 mos · Hyderabad, Telangana, India

Development and Integration of ML models in test:
Developed and integrated machine learning models to improve test automation by 5 hours. The first model used clustering algorithms to set the priority of test cases.
The second model used NLP to classify failed test cases is driver failure or not by analysing the failed summary file.
The last model used LSTM to generate description for describing the failure and raising a ticket.
◦ Data Pipelining: Moving data between different stages of an ML workflow to automate and streamline these processes, making it easier to iterate on different models and improve overall performance

Data PipelinesMachine LearningPython (Programming Language)Data ProcessingPredictive ModelingTensorFlow+16