Embedded Parallelization Tooling
Yocto Integration for Multi-Core Analysis
Yocto/BitBake integration for embedded multicore analysis
8x
Estimated Speedup
Analyzer-estimated upper bound on 8-core Exynos platform
6
Architectures
ARM, MIPS, PowerPC, x86 support
4
Demo Apps
Ported to Yocto build system
The Problem
Demanding embedded applications (autonomous driving, 5G, avionics) require multi-core platforms, but there's no "multi-core compiler." Developers spent significant time on manual parallelization analysis. The analysis tools couldn't process applications built with Yocto Project - blocking entry into the AUTOSAR Adaptive Platform market for autonomous vehicles.
The Solution
Developed custom Yocto layers (meta-slx-toolchain, meta-slx-demos) that integrate instrumentation into the BitBake build process. Created toolchain wrappers that intercept compilation to instrument binaries for parallelization analysis. Demonstrated end-to-end AUTOSAR Adaptive application analysis for an autonomous driving use case including lane detection, vehicle detection, and emergency braking.
Tech Stack
Build Systems
Compilers
Languages
Platforms
Automotive
My Role: Software Engineering Intern
- Designed and implemented meta-slx-toolchain Yocto layer architecture
- Built toolchain wrapper scripts handling cross-compilation for 6 architectures
- Developed linker wrapper for profiling library integration
- Created remote trace collection scripts for embedded target analysis
- Ported 4 demo applications to Yocto build system
- Demonstrated full AUTOSAR Adaptive application analysis for autonomous driving
Key Differentiators
First Integration: Enabled analysis tools to process Yocto-built applications
Multi-Architecture: ARM, MIPS, PowerPC, x86 - all major embedded processor families
AUTOSAR Adaptive: Cutting-edge automotive standard integration (pre-finalization)
Want to discuss this experience?
I am open to full-time Senior AI/ML Platform Engineer roles where this kind of production AI, data, and platform work is useful.
Other Projects
Enterprise IoT Platform
Multi-Tenant Workflow Automation Infrastructure
Reduced customer onboarding time by 70% and eliminated 40+ hours/month of manual DevOps while the architecture was designed for 1000+ isolated workflow instances.
Zeitgaist
Cross-Lingual Social Intelligence
Synthesizes current insights from 6 platforms in seconds vs hours of manual monitoring, with cross-lingual search and filterable source context.