Aishwarya Nair



Email: anair [AT_SYM] umass [DOT_SYM] edu

Grad student @ UMass Amherst looking to mainly combine Artificial Intelligence and other areas of computer science, like UI/UX design, cloud and IOT and quantum computing. Connect with me on Linkedin and github below!

About me

Work Experience

Data Science Intern and Co-op

Boehringer Ingelheim
May 2024 - present
  • Engineered a Streamlit application to provide a comprehensive solution for deploying image processing models as a service, enhancing accessibility and user experience.
  • Developed and implemented modules for creating, triggering and monitoring Databricks pipelines, reducing latency by up to 60 seconds, and delivering inference results directly from AWS S3 buckets
  • Documented the process of triggering and monitoring databricks pipelines on Confluence increasing development speed and efficiency for related use cases.
  • Created scripts to deploy MMCV and MMDet models on Databricks as serving endpoints using MLflow and Workspace Model Registry for the pilot phase.
  • Enhanced a web scraping algorithm using BeautifulSoup, improving data retrieval efficiency from vendor websites by 50%.
  • Developed a GPT-4-based framework to classify vendors into nine categories based on services offered, including retail pharmacy, retail chain pharmacy, mail-order pharmacy, etc.
  • Modified a GPT-4 prompt to improve classification accuracy by 1% by modification of data attributes.
  • Joining as a fall co-op at BI from September 3 onwards.

AI/ML Engineer

Heystack
September 2021 - August 2023
  • Created a data collection python algorithm to obtain data from major e-commerce websites.
  • Researched emotional response by humans on consumption of snacks and compiled a list of 23 emotions along with their definition to identify presence of these 23 emotions from text.
  • Led a team of 5 for organization of exhaustive list of emotions evoked by snacks.
  • Designed a lexical framework of 22 different emotions expressed by consumers of snacks for PepsiCo using PyTorch, Python and Google Colab. These 22 emotions include reminiscence, amazement, sluggish and more.
  • Supervised a team of 12 for data creation and management for the project to identify emotions in text evoked from snacks.
  • Achieved ~80% accuracy on MPNet models fine tuned for text classification using Tensorflow, Python and Google Colab.
  • Achieved an average of 80% accuracy on fine tuned MPNet models to classify text based on actionable topics for Fortune 500 brands. These topics include parameters related to clothing like sizing, design, sleeve length and footwear like soles, foot support, etc.
  • Deployed fine tuned models on Amazon Web Services (AWS) as REST API using Flask and Python.
  • Developed an algorithm to separate out clauses from sentences using morphological analysis and dependency trees using Python, Google Colab and Google Cloud Natural Language API.
  • Developed a prompt based framework to use GPT models to automate the ontology designed process resulting in faster deployment of vertical based analytics.
  • Deployed an API to train and test GPT models to improve accuracy of the pre-existing models by 2% on average, resulting in more robust predictions and faster metric generation for clients.

Education

University of Massachusetts Amherst

Manning College of Information and Computer Sciences
Masters in Science in Computer Science
September 2023 - July 2025 (Expected)
GPA: 3.96
  • Course work includes STAT 501 (Methods of Applied Statistics), COMPSCI 589 (Machine Learning) and COMPSCI 684 (Reinforcement Learning)

V.E.S. Institute of Technology (VESIT), Mumbai

Bachelor of Engineering - Information Technology
July 2017 - July 2021
CGPA: 9.14/10
  • Stood top 10 in the department 3 times.
  • Course work includes Artifical Intelligence, Data Mining & Business Intelligence and Big Data Analytics.

SMPRS Junior College, Mumbai

Higher Secondary Curriculum
July 2015 - July 2017
Percentage: 83%
  • Granted a scholarship for standing in the top 1% of the state in board exams.
  • Stood third in the college in board exams.
  • Stood in the top 6% in the country in JEE Mains.

Publications

S. Natekar, S. Patil, A. Nair and S. Roychowdhury, "Forest Fire Prediction using LSTM 2021 2nd International Conference for Emerging Technology (INCET), 2021, pp. 1-5 doi: 10.1109/INCET51464.2021.9456113.

Click to jump to description of the project.

Projects

Forest Fire Prediction Using LSTM

Student team leader
July 2017 - July 2021
Link to repository
  • Recipient of research grant by Microsoft under AI for Earth program worth $15000 USD.
  • Created the proposal to secure the grant from Microsoft
  • Achieved ~85% accuracy on LSTM model with Python, Tensorflow and Azure using data from FIIRMS obtained from NASA for India.
  • Led a team of 5: 3 for data collection & data processing 1 for sub model implementation and 2 for web development, testing and documentation.

ASSIST (Assistive Smart Stick For Independent and Safe Travel)

Team Leader
Link to repository
August 2019 - May 2020
  • One out of the 49 teams selected for implementation. These 49 teams were selected out of a total of 1346 teams.
  • Led a team of 3: 2 for Internet of Things (IOT) and one for web development.
  • Created an algorithm using Mobile-Net base models achieving 83% accuracy in detecting objects in front of the visually impaired user using Tensorflow, OpenCV and Python.
  • Mapped the most inclusive spaces using the image processing algorithm detecting braille, ramps and textured areas, Google Maps API, HTML and PHP.
  • Created a visually impaired friendly web application with speak aloud facility for location tracking
  • Link to image of workflow
  • Link to video demonstration

LCP (Long term Capacity Planning)

Team Leader
September 2019 - April 2020
Link to the final presentation of the project
  • Super Team awardee of Project Deep Blue competition organized by Mastek & Majesco.
  • Achieved 86% accuracy in predicting the water demand of Mumbai for the year 2030 using data mining and machine learning algorithms using Python and SciKit Learn.
  • Led a team of 3: 2 people for data collection and processing 1 for web development, testing and documentation.
  • Created and presented the application to the board of directors at Mastek & Majesco.

Volunteering Experience

Buddy Mentor

Department of Information Technology, V.E.S. Institute of Technology
August 2020 - July 2021
  • Mentored 10 second year students as their buddy mentor
  • Guided students about in house and industrial internships, preparation for hackathons and placements, coursework, etc.
  • Organized meetings between professors and with other mentors with the students to inform them about the available research and internship opportunities.
  • Summarized the minutes for these meetings.

Member of Editorial Committee

Praxis Tech Fest, 2018. V.E.S. Institute of Technology
August 2018 - September 2018
  • Designed flyers and posters for events in the Praxis technical fest
  • Coordinated among the organizing teams to maintain registrations for events.
  • Published articles in the magazine about the events.
  • Publicized about the events and their schedules in the college via social media, in person announcements, etc.

Sports Incharge

Class Council – Department of Information Technology
August 2017 - June 2018
  • Organized sports teams among first year Information Technology students to participate in various sporting events
  • Coordinated among the sports council and the class teams about event information and requirements
  • Trained a team of 8 for sports like Dodge ball, Basketball, etc.
  • Scheduled practice sessions for the participating teams for the department.

Extracurricular Participation & Achievements