The shipped record

Work

Systems I've built that are running in the world right now.

2024 — now·Austin, TX

Apple — Software Engineer

The retail platform behind every Apple Store on the planet, at 99.99% uptime.

  • Designed Spring Boot microservices powering appointment and staffing workflows for all Apple Store locations worldwide, under high-volume concurrent load.
  • Engineered event-driven capacity pipelines on Kafka and Temporal: fault-tolerant workflows with zero data loss under partial infrastructure failure.
  • Own the capacity platform's Oracle/NoSQL schema and data access layer; tuned G1GC and thread pools under sustained load toward a 99.99% uptime SLA on hybrid on-prem/AWS infrastructure.
  • Led production rollout of 20+ containerized services on Kubernetes with canary releases and automated rollback, enabling zero-downtime deploys for globally critical operations.
JavaSpring BootKafkaTemporalKubernetes

2024·Santa Clara, CA

Qualcomm Innovation Center — Software Engineer

A short chapter in Snapdragon wearables, before Apple called.

  • Worked on embedded software for Snapdragon smartwatch chipsets: Android/Linux framework layers, HAL interfaces, and SoC-level architecture for small form-factor devices.
  • Hands-on with embedded Linux debugging and power/performance analysis for next-generation IoT products.
Embedded LinuxAndroid HALSoC

summer 2023·Hillsboro, OR

Skyworks Solutions — Software Intern

Made TensorFlow Lite for Microcontrollers measurably faster, then upstreamed it.

  • Optimized LSTM and convolution kernels in tflite-micro for HPC hardware, a 15.86% efficiency gain.
  • Contributed refinements to the requantization scheme in open-source TensorFlow Lite Micro.
  • Built APIs for audio-metric prediction and benchmarking, integrated into a dockerized MLflow pipeline.
tflite-microC++MLflowAudio ML

2021 — 2022·Bangalore, India

Samsung R&D Institute — Senior Software Engineer

Night Mode I helped architect runs on millions of Galaxy phones.

  • Architected the Multiframe Solution Pipeline for Night Mode, a pipes-and-filters design over CPU thread pools and GPU shared memory, now live on millions of Galaxy devices.
  • Engineered 3A (auto-exposure, auto-focus, white balance), software ISP, Bayer multiframe blending, and deghosting in C++ with Neon SIMD intrinsics on Qualcomm and Exynos SoCs, cutting AI model runtime 49% under real-time constraints.
  • Redesigned the AI abstraction layer and built a new NDK pipeline to deploy quantized models on on-device NPUs, opening the door to model adoption across device generations.
C++Neon SIMDGPUNPUCamera

2020 — 2021·Remote

Alphaline Wealth Advisors — Software Engineer

Owned a financial news product end to end, from inception to adoption.

  • Built real-time ingestion on Kafka streaming finance and RSS feeds from 300+ sources.
  • Cataloged data with AWS Glue on S3 and prepared it for analytics with Spark ETL jobs.
  • Introduced Airflow scheduling, database optimizations, and NLP preprocessing.
KafkaSparkAWS GlueAirflowNLP

2019·India

Deutsche Bank — Intern

A CI/CD platform serving two thousand engineers.

  • Architected a DevOps CI/CD platform used by ~2,000 people, with live data fetching at 0.2s.

dates approximate, carried over from the previous site

DevOpsCI/CD

2019·India

Impel Labs — Intern

First production ML: automated form scanning with YOLO.

  • Built a YOLO-based form scanning pipeline and improved model accuracy by 9.3%.

dates approximate, carried over from the previous site

YOLOCNNComputer Vision

Credentials

  • MS Computer Science, University of Southern California. GPA 3.9 / 4
  • BTech Computer Science, NIT Karnataka. GPA 3.9 / 4
  • Published: DRIP-Caps, capsule networks for sub-retinal fluid segmentation. Medical & Biological Engineering & Computing (Springer, 2021)
  • KCET rank 86 of ~200,000