Suggestions
Sunil Mallya
CTO/Co-Founder Flip AI
Sunil Mallya is a well-established figure in the tech industry, known for his expertise in building domain-specific language models to enhance developers' experiences.
With a robust background in entrepreneurship and leadership in Engineering and Machine Learning (ML), Sunil has made significant contributions to various projects and services during his tenure at Amazon Web Services (AWS) and other notable organizations.
Having led teams and projects such as Amazon Bedrock, Amazon Comprehend, AWS DeepRacer, and AWS ML Lab, Sunil has a proven track record in areas including Reinforcement Learning (RL), Distributed Systems, and Natural Language Processing (NLP). Moreover, he has an impressive portfolio of 29 filed patents spanning ML, RL, NLP, and Distributed Systems.
Sunil's professional journey includes prominent roles such as the Former Head of Engineering (Sr. Software Development Manager) for the NLP service Comprehend at AWS and Principal Scientist at AWS ML Solutions Lab. Noteworthy achievements include being the inventor of AWS DeepRacer and leading healthcare ML/NLP research initiatives on Electronic Health Records (EHR) in collaboration with Cerner.
Prior to his time at AWS, Sunil co-founded Neon Labs, where he spearheaded engineering efforts, expanded the team, and developed technologies for video processing, image content delivery networks (CDN), as well as neuroscience and machine learning-based image interaction algorithms. He has also contributed to the development of highly scalable systems for Zynga, EMC, and NetApp, including the Zynga API Gateway, web servers/proxies, and distributed file systems.
Sunil has a strong educational foundation with a Master's degree in Computer Science from Brown University. He has held key positions such as CTO/Co-Founder at Flip AI, Principal Deep Learning Scientist at AWS DeepRacer, and Co-founder/CTO at Neon Lab. His earlier roles include Senior Software Developer at Zynga and Software Engineer at Dell EMC.
Sunil Mallya is renowned for growing diverse teams, launching innovative ML products, and driving impactful research in the fields of Machine Learning, Reinforcement Learning, and Natural Language Processing. He has showcased his expertise through numerous patents, research publications at premier conferences like NeurIPS and ICRA, and by delivering talks at esteemed industry and academic events.
His passion for solving complex problems that transcend multiple domains and his commitment to continuous learning are highlighted in his professional endeavors, making him an influential figure in the tech community.
Highlights
McKinsey has its take on #GenAI jobs and future of many occupations, But they seems be hyper-focused on number crunching rather than how the world is going to look different. Here's my take specifically w.r.t Software Engineering -
While there's been rapid progress in coding skills, at present they are far from being your team mate or replacing developers. At best they have replaced stackoverflow and search engines. We are a couple of generations of LLM iterations away from LLMs being meaningful in large codebases at which point you can reduce headcount.
What does that mean for developers? Is the job going to be obsolete?
Absolutely NO. As with any automation, we move to the next level of abstraction. Developers will solve harder problems, be more productive, and take on more business critical decision making.
I haven't seen the future, but one of the two scenarios are likely to play-out
a) Far fewer programmers: AI gets so good that we don't need as many coders as we do. There are 27M coders today, we may be not see growth in these numbers. McKinsey seems to indicate higher level of automation adoption in STEM, implying fewer professional humans required in the future.
b) Everything in the word becomes programmable: Throughout history, automation and democratization of technology has up-skilled everyone. Mobile phones changed the interfaces in how we engage with other humans. Similarly given how coding could be an accessible skills to every human with a natural interface, we may evolve into a world where not just computers but every device or interface may become programable. Nothing exists in a silo. Imagine sending an email laying out your weekly meal plan that results in automated shopping and scheduled meal deliveries, your entire family vacation is planned through a single email. In the rural parts, a farmer could send a text message to activate a network of drones that monitor crop health, automatically adjusting irrigation and applying the precise amount of fertilizer or pesticides where needed. And finally imagine, creating your own personal workflows out of the apps on your phone after completing a workout in your fitness app, your phone could automatically log the exercise in a health-tracking app, update your calorie count in a nutrition app, and send a reminder to drink water through your reminder app. Or, upon arriving at a new destination using Maps, your phone could automatically open a travel app to show nearby attractions, restaurants, or hotel options, making your experience more seamless and connected.
The take home message here being, you are no longer restricted by features provided by a product, instead you are able to ship your own personal features that are more meaningful to you.
My vote is on scenario b. Everything becomes connected. Welcome to the programmable world, my new developer.
#LLM #AI #Future #GenAI #software
Enterprises don’t like their data leaving their boundaries, In this blog we describe how Flip AI has built its LLMs that can be deployed in customer's VPC to solve Observability for customers.
Privacy and Security are non-negotiable at an enterprise and it is time that data is no longer sent out to LLMs, instead LLMs should be next to your data.
#Devops #LLM #AI #Observability #SRE #ProductionEngineering #APM #Logs #MELT
