Hi, I'm Ananya Deoghare.
A
Self-driven, quick starter, passionate programmer with a curious mind who enjoys solving a complex and challenging real-world problems.
About
I am a Master’s graduate in Electrical and Computer Engineering from the University of California, Los Angeles (UCLA), with a deep focus on Machine Learning, Artificial Intelligence, and Software Development. Prior to UCLA, I spent two years at Accenture in India, where I applied data-driven solutions to solve real-world business problems across the healthcare and pharma sectors.
Passionate about using technology to drive meaningful impact, I’m fueled by curiosity, creativity, and a relentless pursuit of excellence. I thrive on transforming complex challenges into practical, human-centered solutions that make a difference.
Currently, I work as an Algorithm Engineer at QuinStreet, where I design and deploy machine learning models—particularly Gradient Boosting Machines (GBMs)—to power ranking and predictive analytics in the personal loans and banking domains. My work involves rigorous feature engineering, cross-validation, and model optimization, directly enhancing decision-making, relevance, and business performance.
- Tools & Languages: Git, Python, MATLAB, Java, C, C++, HTML, Visual Studio, JIRA, GCP, AWS, Perforce
- Databases: MySQL, Apache Hive(Data Warehouse), Amazon Redshift
- Libraries: NumPy, Pandas, OpenCV, Scipy, Cython
- Frameworks: Keras, TensorFlow, PyTorch
- Statistics & Machine Learning: Statistical Analysis, Data Mining, Data Visualization, Image and Video Processing, Computer Vision, Clustering and Classification, Deep Learning, Feature Extraction, Signal processing, Reinforcement Learning
I have pursued my hobbies of Bharatnatyam, an Indian classical dance form and cooking thereby helping me stay focussed and composed. I am a compassionate leader who communicates well and is good at managing people coming from different cultures.
I am looking for an opportunity to work in a research/ engineering/ applications team. I can give references who can testify about my ability to deliver excellent results and work very well in a team.
Experience
- Developed a robust ad-tech platform leveraging Python and machine learning pipelines to analyze creative performance, enabling scalable A/B testing and real-time feedback loops.
- Designed and optimized an automated simulation testing suite using Cython, achieving 95% fidelity with production environments and significantly improving test coverage and computational efficiency.
- Engineered high-impact feature sets using domain-specific aggregations, temporal windows, statistical summaries, and one-hot encoding—improving model generalization and reducing overfitting.
- Implemented advanced supervised learning models (e.g., Gradient Boosting Machines, Logistic Regression), using Bayesian optimization and cross-validation to fine-tune hyperparameters and boost ranking accuracy.
- Conducted in-depth feature importance analysis (e.g., SHAP, permutation importance), removing low-signal features and reducing model complexity by 30% without sacrificing performance.
- Delivered production-ready machine learning pipelines in collaboration with product and data engineering teams, improving data-driven decision-making speed by 10% and enhancing metric tracking for advertiser engagement.
- Tools & Techniques: Python, Scikit-learn, Cython, Redshift, MySQL, Feature Engineering, SHAP, A/B Testing, Cross-Validation, GBM, Data Pipelines, AdTech, Ranking Models
-
TA for the course:
- Food Politics in the World of Arts and Culture/Dance Department
- Mathematics for Life Scientists in the Life Sciences Department.
- Introduction to Archaeology in the Anthropology Department
- Evaluated student work and constructively guiding revisions and instructional assistance to students.
- Worked with a highly diverse student population and adapting and improving practices based on career development experiences.
- Organized and oversaw 2 discussion sections and 2 lab sessions for a course of 30 students, where I guidedstudents to get familiar with mathematical modeling in Python.
- Assisted faculty with preparations for the course, including grading and providing feedback on student assignments.
- Helped students get familiar with the course work and teach them.
- Tools: Python, Verbal Communication, Active Listening Skills, Patience
- Spearheaded the full-stack development of an ad-tech analytics platform, integrating end-to-end machine learning pipelines in Python for real-time evaluation of ad creative performance.
- Designed and implemented deep learning models for computer vision tasks—such as object detection and image scoring—resulting in a 25% increase in the accuracy of ad engagement classification.
- Applied supervised learning and statistical modeling techniques to predict user interaction rates, followed by rigorous A/B testing and performance benchmarking, which led to a 5% boost in overall client satisfaction.
- Engineered modular, reusable components for data preprocessing, model training, and evaluation, reducing pipeline execution time and improving development efficiency across ML iterations.
- Collaborated with cross-functional teams (data, product, and engineering) to deploy scalable AI-driven insights into production, contributing to a 15% increase in team efficiency and enhanced business impact through predictive analytics.
- Tools & Techniques: Python, PyTorch, Scikit-learn, Computer Vision, Deep Learning, MySQL, React, Git, SQL, A/B Testing, Feature Engineering, Full-stack Development
- Wrote 2 chapters for a book on Computational Imaging, which was published by MIT Press.
- Collaborated with a team of engineers to develop and implement a state-of-the-art Shift Robust Loss Function for rPPG, resulting in decreased error by 40%
- Worked with team to diagnose skin-tone bias in medical application using multimodal fusion between radar and RBG data. As a result, the team was able to develop an algorithm that improved accuracy by 75%. The same work has been published in SIGGRAPH.
- Tools: Python, OpenCV, Keras, Tensorflow, PyTorch, Computer Vision, Computational Robotics, Scipy
- Qualified as a semi-finalist in the Global Innovation Challenge held by Accenture after demonstrating exceptional problem-solving skills, creativity, and ability to work under pressure.
- Served as a point of contact between client, technical and test teams to ensure smooth communication and progress.
- Developed and implemented innovative data analysis techniques to enhance the accuracy of drug sales performance monitoring, resulting in a 20% increase in competitor match rate and recognition from the Australian team.
- Successfully trained 10 new engineering graduates on the work done in the team, helping them get familiar with the work and contributing to their success.
- Leveraged expertise in ETL and BI tools to represent monthly and weekly effectiveness of client resources, leading to a 15% boost in productivity.
- Developed and implemented agile development methodologies to enhance team productivity, resulting in a 25% increase in project completion rate.
- Spearheaded the creation and implementation of automated data cleaning and processing workflows, resulting in a 50% reduction in analysis time for pharma sales data.
- Tools: Python, SQL, Apache, Hive, Informatica BDM, Qlikview, Data Warehouse, Tableau
- Collaborated with Dr. Sundaram and a team of researchers to pioneer new methods for faster and more accurate diagnosis of Autism, resulting in improved treatment outcomes.
- Successfully cleaned and rearranged ABIDE Dataset (depending on correlation between different regions of the brain) to improve data correlation, leading to more accurate predictions.
- Developed and optimized deep neural networks to analyze fMRI scans for early detection of Autism, achieving an 80% accuracy rate on a large set of cleaned data.
- You can locate the report pertaining to this project here.
- Tools: Python, Pytorch, tensorflow, OpenCV, MATLAB, Research, Scipy, Pandas
Projects
- Tools: Computer Vision, Python, OpenCV, Haar Cascade
- Provide a 10-30 seconds video
- Process the video using Signal processing
- Output is the approximate heart rate of the person in the video
- An error rate of 3-4 heartbeats
An Automatic Garbage Segregator Arm using Raspberry Pi
- Tools: Python, OpenCV, Deep Learning, Trasnfer learning, Pytorch, Tensorflow, Pandas
- Engineered a crane that could segregate waste into biodegradable, non-biodegradable, electronic waste with an accuracy of 95.18%.
- Tested various Feature Extraction techniques like PCA, LDR and Convolutional Neural Networks. Report
Multi-Class EEG Motor Imagery Classification Using Deep Learning Architectures
- Tools: Python, Computer Vision, Deep Learning, OpenCV, Keras, Tensorflow, Pandas
- This project explores deep learning techniques for multi- class classification of EEG signals for motor imagery tasks.
- Various neural network architectures such as CNNs, LSTMs, RNNs, VAEs, Transformers, and attention are implemented and tuned.
- Effects of pre-processing and time duration on accuracy are analyzed. Results from training on every subject and the entire dataset are compared.
- The project shows the potential of deep learning techniques for EEG signal analysis and classification, highlighting the accuracy of the models and their ability to learn underlying patterns in EEG data. Report
Built a small replica of Google Home using various PCBs and Sensors
- Tools: Python, OpenCV, Raspberry-Pi, Sensor tech
- Successfully built a Google Home prototype using Raspberry Pi, demonstrating an understanding of various types of communication methods.
- Acquired new skills in [Raspberry Pi, Google Home] through self-directed learning and experimentation.
Mono-SET: 6-DOF Monocular Slam for Eye-Tracking
- Tools: Python, Computational Imaging, OpenCV, Raspberry-Pi, Sensor tech
- Studied the effectiveness of indoor tracking for gaze estimation using ORB-SLAM using a custom-built eye-gaze wearable setup.
- The gaze of the user is estimated using a modified U-Net architecture to extract pupil parameters from a stationary pose.
- An Inertial Measuring Unit (IMU) and Fish-Eye camera uses the ORB-SLAM Feature Extractor and Indoor SLAM Environment to track the orientation and movement of the user. The gaze prediction using the data from the ORB-SLAM estimation is used to calculate the distance between the user and the gaze point which is further compared with the ground truth.
Shift Robust Loss Function for remote PhotoPlethysmoGraphy (rPPG)
- Tools: Python, OpenCV, Object Detection, Computer Vision, Deep learning
- rPPG is a technique that uses slight color changes in a person’s face to find remote-PPG signal and estimates heart rate from this signal using signal processing techniques.
- We used an RGB camera to obtain face videos of volunteers and a finger pulse oximeter to obtain the ground truth PPG signal while collecting data from rPPG. These two devices are not hardware triggered, so there is a slight lag between the frames obtained from the camera and PPG signal
- State-of-the-art methods require the ground truth data to be exactly aligned with the signals.
- To address this problem, we propose a novel shift-robust loss function for regression that enables the network to learn from misaligned data to achieve better results than recent methods Report
Summarized the paper "The Dynamics of Inequality"
- Tools: Stochastic processes, ito's Calculus, Brownian Motion
- United states has seen a rapid growth in top inequality in the past forty years, i.e. the rate at which super rich are getting richer is more than the rate at which rich are getting richer, due to which the inequality between the two groups is increasing.
- In the two deviations that this paper proposes, they consider two reasons namely, "changes in skill prices" or "rise in superstar entrepreneurs or managers". Report
Skills
Languages and Databases
Python
MATLAB
HTML
MySQL
Shell Scripting
Java
C
R
Libraries
NumPy
Pandas
OpenCV
scikit-learn
Frameworks
Keras
TensorFlow
PyTorch
Platforms
Google Colab
QlikView and QlikSense
Soft Skills
1. Leadership
2. Professional Bharatnatyam Dancer. I have also trained in 10+ dance forms!
3. Communication - Ability to communicate in precise and concise manner.
4. Ability to take a 360 view of the problem and define the problem correctly.
5. I enjoy working with Cross cultural teams, and have worked very well.
Other
Git
AWS
Raspberry Pi
Arduino
Firebird Robot
Education
University of California, Los Angeles
Los Angeles, USA
Degree: Master of Science in Electrical and Computer Engineering
CGPA: 4.0/4.0
- Computer Vision
- Large Scale Data Mining
- Large scale Social and Complex Networks
- Neural Networks and Deep Learning
- Machine Learning and Modeling in Bio Engineering
- Computational Robotics
- Computational Imaging
- Decision Making in Stochastic processes
- Technology Management
- Physics and Informatics of Medical Imaging
Relevant Courseworks:
Bengaluru, India
Degree: Bachelor of Technology in Electronics and Communication Technology
CGPA: 3.69/4
- Introduction to Python, C & Embedded Systems
- Probability and Random processes
- Linear Algebra
- Information Theory and Coding
- Machine Learning
- Digital Signal Processing
- Speech Processing
Relevant Courseworks: