Welcome!

About Me

Having honed my skills with a Masters in Computer Science and Engineering from IIIT Bangalore, my professional trajectory over the past three years has taken me through the nuanced terrains of Telecom, Retail, and Technology. Each domain presented its unique set of challenges. Yet, with a persistent focus on harnessing data-driven insights, I’ve consistently translated challenges into solutions and value propositions.

Presently, I’m deepening my expertise in the MS Analytics program at Georgia Tech. My specialization in Computational Data Analytics has been a rich learning ground. Here, I’ve bolstered my analytical capabilities and delved into pioneering research in Computer Vision. Moreover, I’ve embarked on practical projects, notably one that focuses on the riveting world of Generative AI.

Technical Prowess

Languages & Libraries: Python, Java, R, Pytorch, Scikit-Learn, Pandas, NumPy, Matplotlib, OpenCV, OpenAI Gym, Tensorflow, PyCaret, PySpark
Infrastructure & Databases: MySQL, Apache Kafka, Docker, ELK, AWS EC2, SQL, Google Cloud
Visualization Tools: Tableau, Seaborn, D3.JS, Kepler.GL
Analytical Competencies: Regression, Clustering, Classification, A/B Testing, Deep Learning, Computer Vision
Accolades: IBM Quantum Challenge 2021 - Advanced

As my graduation in December 2023 approaches, I’m poised to embrace opportunities as a Data Scientist, Applied Scientist, or Machine Learning Engineer. Committed to bringing a blend of academic insights and practical industry experience, I am enthusiastic about joining teams that value cutting-edge innovation and robust data-centric strategies.

If you believe in the transformative potential of data as much as I do, and are looking for someone with a proven knack for insight-driven innovation, I’m all ears! 🚀 Please feel free to touch base at spotluri31@gatech.edu or sairakshithp97@gmail.com.

Industry Experience

Knowledge Base Enhancement with Generative AI, Sponsored Practicum Project with NCR Corporation

Aug 2023 – present | Atlanta, United States

Collaborating to enhance NCR’s Knowledge Repository through Retrieval Augmented Generation and Generative AI, using tools like Microsoft Azure, Databricks, Chat-GPT and LangChain.
Utilizing LLM-based embeddings for refined content extraction and integrating vector databases to ensure precise article matching.
Crafting a streamlined QA chatbot that blends Large Language Model capabilities with industry-specific knowledge to significantly elevate customer query resolution.

Data Science Intern, Cognira

May 2023 – Aug 2023 | Atlanta, United States

Orchestrated and executed PySpark pipelines enhancing data integration, leading to a consolidated sales and promotional dataset.
Undertook data analyses to uncover previously unnoticed pantry loading effects spanning 202 top-volume categories.
Employed sophisticated statistical methods and introduced innovative features, significantly elevating the forecasting mechanism for a Tier-1 retail client.
Revamped inventory strategies, setting the stage for potential yearly savings of an impressive $480,000.

SWE - Data Scientist, Reliance Jio AI-COE

Aug 2020 – Jul 2022 | Hyderabad, India

Leveraged PySpark for Extract Transform Load processes, boosting data efficiency by 41%. This paved the way for optimized downstream applications like Customer Segmentation and Large-Scale Subscriber Monitoring for over 400 million telecom subscribers.
Pioneered a Digital Twin Model using LightGBM and XGBoost, achieving a remarkable increase in AUC from 0.61 to 0.82 post meticulous data tuning.
Introduced an SHAP-based explainability mechanism, offering 4G/5G Network Insights which catalyzed significant service enhancements.
Spearheaded collaboration among cross-functional teams, playing a crucial role as a bridge. This synergy was crucial in harnessing generated reason codes for Targeted Promotions, propelling monetization efforts.

Research

AI Forest: Cognition in the Wild

Supervised by: Dr. Jacob Abernethy, College of Computing, Georgia Tech & Dr. Marcela Benítez, Department of Anthropology, Emory University

Deployed YOLOv7 for primate identification within the AI Forest initiative, witnessing a substantial boost in accuracy rates — from 0.85 to 0.92, specifically for wild capuchin images.
The ensuing phase is set to witness real-time deployment within a wildlife sanctuary located in Costa Rica, geared towards recognizing and tracking capuchin monkeys of identical species.
This pioneering project has been honored with the esteemed AI.Humanity seed grant bestowed by both Georgia Tech and Emory University.

Generation Of Realistic Vehicle Trajectories From Video Streams

Collaboration: Research Co-op with Siemens Technology

Researched multiple behavioral models for autonomous navigation using Reinforcement (RL) and Imitation Learning (IL).
Finalized the ‘Learning by Cheating’ method, implementing it through Tensorflow, OpenAI Gym, and CARLA.
Devised a pipeline that leverages Mask-RCNN and IOU Tracking to seamlessly transform raw traffic camera feeds into distinct vehicle trajectories. With this strategy, we achieved a 100% surge in our trajectory dataset.

Agent-Based Modeling of Data Localization

This agent-based model simulatees the trade of digital services between consumers and producers. An augmented gravity trade model with various data regulatory constraints accounted for costs of the services. Additionally, The work was published as a dedicated chapter in Data-centric Living: Algorithms, Digitization and Regulation and was also cited by United Nations Digital Economy Report 2021.

Projects

Telecommunication Spectrum Data Visualization

Project Overview:
- Visualized dense Telecom Spectrum data employing techniques like Heatmap, Treemap, and Geographical Map Visualization, integrating them onto the website.
- Implemented visualization using D3.js and JavaScript.
Technical Stack & Processes:
- Development Tools: Utilized D3.js and JavaScript for data visualization.
- DevOps & Deployment: Leveraged Jenkins, Docker, ELK, and AWS during the portal’s development phase. Hosted the final product using Netlify.
Live Demo: View Website

Agent Based Modeling and Simulation Of Drug-Resistant Diseases

This agent-based model was developed to study the onset of drug resistance in Mycobacterium tuberculosis. The model uses RepastCity to simulate human movements on GIS-based urban enviroments and simulates subsequent spread of the disease. Our model was able to implicate interesting findings such as differences between drug-resistant and drug-susceptible diseases.

Image Style Transfer Using Convolutional Neural Networks

Implementations of neural style transfer(NST) were developed to compose images in the style of another image. Different versions of the NST are implemented, such as the VGG19-based single style transfer and Segmentation style transfer