Jun 2022 – Oct 2025 · 3 yrs 4 mos · San Francisco, California, United States · Remote

• Software Architect/UTL for the Abnormal Data Platform - leading 20+ engineers across multiple teams in US/Singapore/India to deliver high-scale, low latency systems for data ingestion, feature stores and knowledge layers serving ML detection systems.

Nov 2021 – Jun 2022 · 7 mos · San Francisco Bay Area

Nov 2016 – Oct 2021 · 4 yrs 11 mos · San Francisco Bay Area

Jan 2015 – Oct 2016 · 1 yr 9 mos · San Francisco Bay Area

Nov 2014 – Dec 2014 · 1 mo

• Owned crucial Software-in-Silicon features for the M7 processor in SPARC Functional Simulator (C++).
• Acted as SME on the DAX co-processor, found bugs in RTL implementation and test framework.
• Added debugging/tracing/test-generation/snapshot infrastructure for M7.
• Proactively initiated the validation effort for Solaris-12 and built out missing simulator features.
• Wrote automated C++ class generator, saving months of development time.

Jan 2012 – Oct 2014 · 2 yrs 9 mos

• Member of the Functional Simulation team which models SPARC server systems.
• Key responsibilities include:
• 1. Modeling new Sparc cores and devices like NICs, run firmware stack on top and generate traces
• 2. Co-simulation with RTL for functional validation
• 3. Automated test development and regressions
• 4. Faithful representation of hardware to enable pre-Silicon software development and debugging

May 2011 – Aug 2011 · 3 mos

• Understood the architecture of a coherency ASIC and identified key features for initial testing. Developed Perl-scripts which used JTAG debug interface to communicate with hardware. This helped significantly in post-Si validation of the ASIC.

Sep 2010 – May 2011 · 8 mos

• Built a LAMP(Linux, Apache, MySQL, Perl) system from scratch along with a Perl-based web-crawler which downloads relevant data from the public forum website and updates the database.
• Created a template implementation which can process almost any forum based on HTML tags.
• Identified underlying social network properties through tools like ORA, Network-Bench.
• Downloaded 100+ GB raw HTML pages and created 10 million rows table using Perl.
• Modular design enabled reuse of the code-base for multiple forums, furthering research in vernacular web.

Aug 2008 – Jun 2010 · 1 yr 10 mos

• 1. Implemented Advanced Image Processing algorithms, optimized to run in real-time, on OMAP-4 mobile computing platform.
• 2. Spatial and Temporal Noise filtering, Distortion correction, Face detection and Contrast Enhancement; using Imaging accelerators on SoC.
• 3. Co-developed a software framework for abstracting complexity of programming hardware accelerators and make them more accessible to customers for customization. This framework is capable of efficient scheduling of imaging accelerators, resource management and self-configuration.
• 4. Validated functionally through CCS Simulator, FPGA Emulator and Silicon samples by writing extensive test-suite, with automated reference generation and comparison.
• 5. Defined and programmed APIs for IMX co-processor simplifying imaging algorithms.