Ravi Hastantram

CTO

Bengaluru, Karnataka, India19 yrs 9 mos experience

Most Likely To SwitchAI Enabled

Key Highlights

Led cross-geo teams in developing edge algorithms.
Architected next-gen image delivery systems at web scale.
Optimized machine learning techniques for image processing.

Stackforce AI infers this person is a SaaS expert with a focus on machine learning and edge computing.

Contact

Skills

Core Skills

Large Language Models (llm)Engineering ManagementSoftware ManagementMachine LearningDeep LearningEdge ComputingImage Processing

Other Skills

Responsible AITransformer ModelsGenerative AICross-functional Team LeadershipHigh Performance TeamsComputer VisionC++Android DevelopmentEmbedded SoftwareGPUMultithreadingDebuggingDevice DriversOpenCLOpenGL

About

Leading a cross geo team to develop edge algorithms & framework to run Deep learning models (CNNs)/vision and sensor based algorithms on device. Web scale image delivery, Web performance. Adaptive image optimizations based on Machine learning models to deliver best user experience. Architecting the next gen Image delivery at web scale to effectively use all hardware resources on a platform. High performance multi threaded runtime systems for heterogeneous programming on Android.

Experience

19 yrs 9 mos

Total Experience

3 yrs 3 mos

Average Tenure

5 yrs 6 mos

Current Experience

Amazon

2 roles

Software Development Manager

Jul 2024 – Present · 1 yr 9 mos

Excited to serve a team of talented engineers at Amazon Photos in Bengaluru, building the team from scratch. Infusing AI especially multimodal LLMs into all the experiences customers love like natural language search, semantic understanding, curation & presentation of rich memories. Started my journey at Amazon Bengaluru helping build the graphics pieces for upcoming amazon devices.

Large Language Models (LLM)Responsible AITransformer ModelsGenerative AICross-functional Team LeadershipHigh Performance Teams+2

Software Development Manager

Sep 2020 – Jun 2024 · 3 yrs 9 mos

Amazon Textract is a deep learning-based document analysis service that allows customers to automatically extract business-critical data from images/ documents. Serving a team of talented engineers to develop next gen features using the power of LLMs along these vectors 1/ Develop and optimize inference serving systems for responsible use of large models like Claude and in-house supervised fine-tuned models. 2/ New architectures for smart orchestration of different large language models 3/ Innovate on scaling strategies to optimize cost and latency of large model inferences 4/ Enhance security of model hosting platforms and execution of generated code. 5/ Continuously monitor and evaluate large model accuracy through rigorous testing methodologies 6/ Drive optimizations around predictive modeling, caching, compression, and other techniques to minimize large model inference cost
AWS SageMaker Edge/AWS Lookout for Vision enables machine learning on edge devices by optimizing, securing, and deploying models to the edge, and then monitoring these models on your fleet of devices, such as smart cameras, robots, and other smart-electronics, to reduce ongoing operational cost. Leading a team of talented engineers to democratize Edge ML for all customers across different verticals.

Large Language Models (LLM)Responsible AITransformer ModelsGenerative AICross-functional Team LeadershipHigh Performance Teams+2

Nauto

2 roles

Senior Engineering Manager

Jan 2020 – Sep 2020 · 8 mos

Edge Platform Team

Engineering Manager/Lead, Edge Platform Team

May 2018 – Dec 2019 · 1 yr 7 mos

Leading a cross geo team to develop edge algorithms & framework to run Deep learning models (CNNs)/vision and sensor based algorithms on device. Efficient scheduling to achieve low latency inference and high throughput utilizing all the resources on the platform (CPU/GPU/DSP) in a power constrained environment. In depth device performance analysis to understand and optimize the resource & power usage.
Combining leading edge vehicle hardware with an ever-learning artificial intelligence platform,
Nauto prevents accidents and saves lives. Nauto is funded by Greylock Partners, Softbank and Playground Global.

Instart logic

Staff Software Engineer

Aug 2015 – Jan 2018 · 2 yrs 5 mos · San Francisco Bay Area

Optimized our unique Machine learning based image optimization technique by 7x using separable convolutions and vectorized SSE instructions.
Designed and implemented a new architecture for delivery and optimization of images which unlocked new use cases and increased agility of feature releases.
Architected and implemented a novel methodology to leverage video codec techniques for image optimizations.
Designed a potential video delivery solution to optimize both streaming formats like HLS, MPEG-DASH and progressive videos with video codecs like H.264/VP9.

Qualcomm

Senior Software Engineer

Aug 2013 – Aug 2015 · 2 yrs · San Francisco Bay Area

Parallel runtime, programming model to exploit heterogeneity in Qualcomm SOCs (Qualcomm Symphony System Manager SDK, MARE), it's a task based runtime which relieves the programmer from the complexities of multi threading. I am leading the effort to add GPU to the mix, so that a uniform task based programming model is exposed to the user. All the data movements are managed transparently by the runtime. I extensively use c++11 features, template meta programming, policy based design patterns & other advanced c++11 mechanics to achieve efficient execution of a dynamic task graph with dependencies.

Amd

MTS

Jan 2012 – Aug 2013 · 1 yr 7 mos · Austin, Texas Area

I am currently part of the Heterogeneous Systems Architecture (HSA) team responsible for developing the runtime to expose the compute capabilities of the graphics hardware; i am currently focusing on topology discovery, reporting and leading the effort on inter operability of compute stack (HSA) with DirectX and OpenGL. Also enabling existing programming models like OpenCL and new models like C++ AMP to benefit from the HSA.

Intel corporation

Software Engineer

Jan 2005 – Jan 2011 · 6 yrs · Bengaluru Area, India

As a software engineer at Intel i've worked on different aspects of Graphics driver software like HDMI, HDCP protocol, ACPI, power management etc. I started my career at Intel in the Tools team implementing the core modules used by different work flow management & automation tools. I also had a short stint as an Applications Engineer working with key customers to enable their launches;