Sorry, unable to follow Kaveesh. Please try again.
Hi, I'm Kaveesh Khattar. 👋 I'm currently pursuing an MS in Computer Science at NYU and exploring ML systems. I previously worked as a Software Engineer at AT&T on the Network API Platform team, building cloud-native infrastructure for APIs that abstract 5G network capabilities. My interests lie at the intersection of AI infrastructure, distributed systems, Kubernetes, and cloud-native platforms. I'm fascinated by the engineering challenges behind serving large language models in production, making inference faster, more reliable, and more scalable. At AT&T, I worked on cloud infrastructure, platform engineering, and observability using Java, Kubernetes, Azure, OpenTelemetry, Prometheus, Grafana, Terraform, Helm, and CI/CD to improve the reliability and scalability of production services. Beyond engineering, I've conducted research in NLP, computer vision, and generative AI, resulting in publications in Springer and SCOPUS-indexed venues. I also hold a US patent for applying machine learning to optimize distributed network performance. I'm always excited to collaborate on cloud-native systems, Kubernetes, distributed systems, and AI infrastructure. Let's connect!
Worked on the Network API Platform team, building cloud-native infrastructure for APIs that abstract 5G network capabilities.
Worked on making intern onboarding seamless and fully self-guided. Streamlined onboarding for 50+ interns by developing a full-stack platform (React.js, Spring Boot, PostgreSQL, Docker) that reduced sync-up time through progress tracking, forums for resolving technical roadblocks, and automating logistics like access approvals and room bookings.
Worked on predicting chronic inflammation from physiological data. Researched and created a driver wellness tool, using pandas, NumPy and scikit-learn, achieving 85% accuracy in detecting chronic inflammation by constructing a dataset with 5,000 samples and markers like Heart Rate Variability (HRV), skin impedance, and temperature, and employing traditional and deep learning algorithms for predictions. Created a Tkinter UI Dashboard using Agile methodology, reducing presentation prep time by 2 days and boosting investor engagement.
Worked on modeling the psychology behind disinformation under Dr. Bhaskarjyoti Das. Investigated the underlying psychological drivers behind the production of disinformation through machine learning (Naive Bayes, SVM, XGBoost) with TFIDF Vectorization & deep learning methods (DistilBERT, spaCy with bert-base-cased), achieving 82% accuracy. Enhanced model consistency with kTrain-DistilBERT and interpretability with LIME, achieving 99.2% confidence indicating high reliability, and making the model more understandable.
Organised CS workshops averaging 60+ attendees and an App Development seminar with 75+ participants averaging 90% engagement.
Started & grew a CS club through outreach & workshops. Led promotion of club objectives in the technical domain and landed 50+ new sign-ups.