Personal AI enables companies to build high-performing virtual employees called “Personas.” Each Persona is specialized in a specific area, carries deep expertise, and plays a key role within the organization.

Designed for seamless human-AI collaboration, Personas work much like human colleagues- researching, completing tasks independently, and integrating with other systems to reach goals. They can be trained, deployed, and securely managed through the Personal AI platform.

https://www.personal.ai

Engineered Pax’s Portfolio Manager(PPM) infra from scratch: secure, massive-scale, always-on. Real-time payments, thousands of terminals — handled.

Part of Intel’s DCAI Hyperscaler team

• Worked on implementing Intel optimizations within the AWS Encryption SDK C by leveraging Intel's advanced hardware accelerators like IAA and QAT to enhance encryption performance and efficiency.

• Showed 2x improvement in QEMU VM live migration tests with minimal downtime using VFIO framework on Intel's Sapphire Rapids with Intel's HW Accelerators like QAT and IAA

• Collaborated on the development of kernel patches involving Google's MGLRU and parallel zswap supplemented with Intel hardware accelerators, aimed at offloading memory tiering to Intel hardware and thereby mitigating CPU utilization.

• Facilitated the integration of Intel’s Spark-based optimizations (based on Intel’s OneAPI libraries) into Google Cloud's Dataproc platform. Collaborated with GCP's third-party customers to enable the utilization of the optimizations tailored to enhance the performance of their products.

• Perform workload characterization and profiling on industrial workloads to record nX performance improvement with Intel HW/SW on cloud compute platforms.

• Optimized Intel's multi-kernel OS (mOS) for the Aurora Supercomputer, enhancing performance for biomedical deep learning workloads.

• Implemented thread-aware scheduling techniques, significantly improving CPU performance.

• Utilized advanced Slurm job scheduling and performance tuning, achieving a notable speedup compared to standard Linux distributions.

Teaching Assistant for CSE 110: Principles of Programming(Java). Conducted lab sessions and TA office hours.