Chandra Kumar Ramasamy

Product Manager

Bengaluru, Karnataka, India21 yrs 2 mos experience
Highly Stable

Key Highlights

  • 14 years of experience in embedded systems and DNN development.
  • Led development of custom software frameworks for deep learning.
  • Expert in neural network compression techniques and DNN frameworks.
Stackforce AI infers this person is a specialist in Embedded Systems and Deep Learning technologies.

Contact

Skills

Core Skills

Deep Neural NetworksDnn Performance AccelerationNeural Network CompressionDnn Framework DevelopmentDnn Inference Library DevelopmentIntegrated Development Environment DevelopmentGdb PortingReconfigurable ComputingGnu Tool Chain MaintenanceRisc Architecture SupportGcc Compiler DevelopmentDsp Compiler OptimizationDsp Compiler DevelopmentBackend DevelopmentC Compiler TestingRegression Testing

Other Skills

Deep LearningEmbedded SystemsSoftware DevelopmentProfilerEmbedded SoftwareSoftware EngineeringProgramming LanguagesDebuggingCompilersGNU DebuggerCC/C++ STLSimulationsTestingUSB

About

1. Totally ~14 years of experience in embedded system based projects involving research, design and development of high performance Deep Neural Networks (DNN) platform and system software tools (compiler, assembly to assembly translator, debugger, simulator, profiler and IDE) for RISC, CISC, DSP and Reconfigurable architectures 2. Played the key role in the development of custom light weight software framework for Deep Learning used in Mobile-Flagship/mid-tier/low-end mobile devices 3. Good understanding of commonly used DNNs (AlexNet, GoogLeNet, ResNet, VGGNet, SqueezeNet, Selfie Outfocus, MultiROI detector, LRCN) 4. Strong understanding of Deep Learning Frameworks such as Caffe, TensorFlow and SDNN 5. Lead Tool-chain (debugger, profiler and IDE) development for Samsung Reconfigurable Processor which got commercialized in DTVs 6. Hands-on experience in open source software - LLDB, Eclipse, GCC and GDB 7. Rich experience in project management and process (AGILE) facilitation Currently I am into the research on Neural Network Compression Techniques - network pruning, quantization in particular.

Experience

21 yrs 2 mos
Total Experience
5 yrs 3 mos
Average Tenure
--
Current Experience

Amd

3 roles

Sr. Manager Software System Design

Promoted

Jul 2024Feb 2026 · 1 yr 7 mos

SMTS Software System Design Eng.

Jul 2021Jun 2024 · 2 yrs 11 mos

  • • DNN performance acceleration on CPUs
Deep Neural NetworksDNN Performance Acceleration

MTS Software System Design Eng.

Sep 2018Jun 2021 · 2 yrs 9 mos

  • Research on Neural Network Compression Techniques (network pruning, quantization)
  • DNN Framework and libraries development for CPU Inference
Neural Network CompressionDNN Framework Development

Samsung electronics

2 roles

Chief Engineer

Promoted

Apr 2012Sep 2018 · 6 yrs 5 mos · Bengaluru, Karnataka, India

  • Designed and developed DNN inference library for Samsung Neural network Processor (SNP)
  • Designed and developed Samsung Convolutional Deep Neural Network Model Visualizer using Eclipse RCP
  • Designed and developed Profiler for Samsung Reconfigurable Processor (SRP)
  • Developed Integrated Development Environment (IDE) for SRP using Eclipse CDT
  • Ported LLDB Debugger for SRP
DNN Inference Library DevelopmentIntegrated Development Environment Development

Technical Lead

Jul 2011Apr 2012 · 9 mos · Bengaluru, Karnataka, India

  • • Ported GDB for Samsung Reconfigurable Processors (SRP)
GDB PortingReconfigurable Computing

Ami india (american megatrends india) private limited

Technologist

Apr 2010Jul 2011 · 1 yr 3 mos · Greater Chennai Area

  • Maintenance of existing GNU tool chain for a RISC processor (ABACUS)
  • Added support for a new architecture in GNU tool chain
  • Extensive testing of ported GNU tool chains
GNU Tool Chain MaintenanceRISC Architecture Support

Acme technologies

3 roles

Team Leader

Promoted

Oct 2007Mar 2010 · 2 yrs 5 mos

  • GCC Prototype Compiler for a VLIW DSP:
  • Enabled auto-vectorization in GCC for porting it to 16bit VLIW DSP architecture.
  • Framed GCC training materials. Framed theoretical and practical exercises for trainees.
  • Trained members on GCC compiler basics and on GCC porting. The members were able to get into porting work within the training period of 4 calendar months.
  • Assembly Translator for CISC architectures:
  • Contributed to the overall design of assembly translator.
  • Implemented "target translator" (programming language C++) for translating the assembly of two different CISC architectures to the assembly of a new CISC architecture. This was a prototype implementation. Based on the prototype, the customer awarded the project.
GCC Compiler DevelopmentDSP Compiler Optimization

Senior Software Engineer

May 2005Sep 2007 · 2 yrs 4 mos

  • DSP-C Cross Compiler for iCore3:
  • Worked in backend development (programming language C) and maintenance of 'DSP-C' compiler (for 16 bit DSP).
  • Implemented pipeline hazards optimization for 16 bit DSP compiler. Customer changed the pipeline hazards specification after the implementation. But the change was incorporated into the implementation without any design change. Such was the design made initially.
  • Contributed to the implementation of Instruction Level Parallelism (ILP)
  • Implemented the interface between Instruction Level Parallelism (ILP) and Pipeline hazards optimization.
  • Implemented the Multiplication and Accumulation (MAC) module.
DSP Compiler DevelopmentBackend Development

Software Engineer

Jul 2004Apr 2005 · 9 mos

  • C Cross Compiler for U8 core:
  • Worked in black box testing of 'C' compiler for 8 bit RISC microcontroller.
  • Wrote test specifications for data segment pragma and optimization pragma.
  • Derived minimized decision tables for the test specification.
  • Wrote black box tests based on the minimized decision tables.
  • Used AQVS (acme quality validation system) for regression testing.
  • Validated assembly differences - regression output validation.
  • Ran tests using simulator and emulator. Verified output logs.
  • Reviewed design specification of modules under development.
C Compiler TestingRegression Testing

Education

Birla Institute of Technology and Science, Pilani

M.Tech. — Software Systems

Jan 2015Jan 2017

Anna University Chennai

MBA

Jan 2007Jan 2009

Manonmaniam Sundaranar University

BE — Computer science and Engineering

Jan 2000Jan 2004

Stackforce found 100+ more professionals with Deep Neural Networks & Dnn Performance Acceleration

Explore similar profiles based on matching skills and experience