Think Circuits Engineers

I'll be speaking at the Embedded Vision Summit in San Jose this May, introducing an overview of Deep Neural Network (DNN) training for AI vision applications. This conference is a great place for practicing engineers and innovative executives to catch up on the latest applications and tools for embedded vision and AI. Last year there were some great talks on Vision Language Models (VLM), excited to see what's in store this year! Comment below for a 25% off discount registration code! #embeddedvision

Large IT projects, on average, go 47% over budget and provide 56% less benefit than original planned! This article from 2012 highlights a phenomena that I personally experienced throughout my career and it led me down a path of learning project management and what leads to effective teams and technology development. The four points highlighted in this article are core to how we work at Think Circuits where we strive to execute technology projects that line-up with business value above all else. 1. Managing stakeholders and aligning them with the strategy: We always start by understanding the bigger picture of what our clients are trying to achieve, this ensures our day-to-day decision making is aligned with the overall strategy. 2. Mastering technology and contents: The article estimates that hiring "the appropriate experts can raise performance by as much as 100 percent through their judgement and ability to interpret data patterns." In addition to the decades of experience in our internal team, we maintain relationships with outside collaborators who are at the top of their respective disciplines. 3. Building effective teams: Collaboration is a stated value of our company, ensuring our people have a "common vision, shared team processes, and a high-performance culture" as opposed to "individual work-stream goals." 4. Excelling at project management practices, ensuring even the most complex projects with many moving pieces are smoothly executed, with agile downstream and upstream communication of status and priorities between our clients and contributors. https://lnkd.in/gGXtgGxK

Our thoughts are with our neighbors, friends, and firefighters in SoCal as at least 5 deaths, 2000 buildings destroyed and over $8B in damages due to several uncontrolled wildfires devastating neighborhoods in LA County. Though I am grateful we have had wet weather in the Bay Area, the affected neighborhoods are not that different from where I and thousands of others live, work and go to school. Like many of our neighbors, we are on the California FAIR plan because private insurers will not insure us. This is partly what has driven me to invest my time, talent, and over $80,000 from our company towards coming up with a solution. **There's no such thing as a natural disaster** This is something my wife learned while studying Geography at UC Berkeley. The fact is, we know the climate is changing and the conditions for destructive wildfires are becoming a regular part of life in many areas. Our current man-powered solutions, though heroic, cannot keep up with this new reality. Insurance, which should be a last resort is also not designed or equipped to handle losses of this magnitude and frequency. There's no excuse for willful ignorance of the mounting problem, we must invest in bold proactive solutions using all of the recent advances in UAS, robotics, and AI to create a new standard of infrastructure protecting our communities. Last year at the #XPRIZE #wildfire Team Summit, I presented a motivating example of how we could protect the areas of LA and San Diego with robotic detect and suppress technology. If we can achieve the XPRIZE (prevent fires from spreading in a 1000km^2 area) goals 29 times, then we can protect tens of millions of lives and trillions of dollars of property. In pursuit of these goals our team has been working tirelessly on developing the system: * We developed "Flamecam" a sensor specifically to detect wildfire at high specificity at kilometer distances. We have demonstrated detecting a 2-inch flame at a half-mile distance and are now scaling up the experiments. * We have built and flight-tested small scale model of the suppression UAS and are iterating on the suppression deployment system and extinguishing behavior AI. * We are building a full-scale model of the detection VTOL UAS which will host the Flamecam sensor at a high altitude to quickly detect fire-starts over acres of covered area. * We are in the proposal stage with a major California company to pilot our technology, and we have been invited to submit an NSF SBIR proposal this March. Our most pressing constraint at this point is having access to land to develop and demonstrate our solution. We would love to collaborate with any private landowners or public agencies to allow us access. If anyone has connections to someone that could help us in this regard, please reach out or send them my email [email protected].

We are thrilled to welcome Micah Morales to the team! Micah is an electrical engineering graduate from UC Santa Barbara. He has experience supporting products throughout the hardware design lifecycle from board design to field failure analysis. Previously he has worked at Cisco and Apple, supporting?products like the Apple Watch and several Cisco 8000 systems. His proficiencies include signal integrity and PCB design for high-speed systems.

The first of many fires to be suppressed by our suppression UAS in development to win the #XPRIZE Wildfire competition! Next step is to upgrade the pump to get a more accurate and voluminous water spray.

First flight test of our AIDEN suppression UAS for the #xprize wildfire competition. Soon we will be testing its ability to put out small controlled fires and see how this compares to our simulations.

I have been experimenting with different AI strategies for suppression of small wildfires using limited water payloads for the XPRIZE #wildfire competition. On the top is a "direct attack" strategy where a UAS attempts to put out the active fire region, whereas on the bottom, the UAS deploys its entire water load to create a wet-line fire block in front of the fire. The direct attack UAS on the other hand, manages to put out this fire with only 1.3kg of water! It looks like direct attack on the fire can be more effective than trying to encircle the fire, though this may not hold true for all weather, fuel, and terrain conditions. That said, if direct attack is really more effective for small fires, then this is another advantage of unmanned aerial systems which can be deployed in more hazardous situations than humans to quickly address incipient fires or flare-ups from a larger fire!

Day 3 of 4 at the XPRIZE Wildfire Summit with Chris Sarantos and Michael Pihulic, and its been really interesting to hear about the past, present and future of wildfire tech from both veterans and new entrants. I think with this many people in one place, focused on a singular mission, I feel like this is the beginning of a needed shift in the technology and markets around wildfire, and I imagine the conversations we are having here will seed the next decade of bold innovation towards wildfire and human-robot-nature coexistence. In 10-20 years from now, drones and other robots will be another trusted piece of infrastructure similar to fire hydrants, engines and aircraft. I can't imagine this not happening as the costs to society are staggering (2-4% of US GDP annually and getting worse), and our current systems of fire mitigation, fire response, and insurance are not keeping up. From the inspiring work I've heard from other teams, the technology itself is within reach, and is equally or more beset by economic and regulatory hurdles. I've made it my mission this week to dive into the economic part of the equation and understand what is the best way connect together the value chain from small startups to the trillions of dollars of economic benefits that would be realized if this were developed globally. There are countless groups that feel the pain from increasingly severe wildfires: Fire agencies, homeowners, small business owners, large property owners, insurance companies, utilities, governments, and more. Insurance is particularly interesting because they are urgently in need of a solution in areas like California where many large insurers are pulling out of the market and losing customers. As a group they may have the most immediately realizable benefit if a product can provably move the needle on their risk profiles. It may just be that we as technologists need to speak the right language and go after the right objectives to present the right value proposition to this industry. I thought I'd share these thoughts with my LI community as its becoming clear that this is a situation where a rising tide lifts all boats. Opening this market with innovative business models and raising awareness will be a benefit to all the XPRIZE teams and the rest of the fire community. One more day to go, then back home to put our ideas to practice! tcwildfire.com

Getting ready to present at the XPRIZE Wildfire Team summit this week in Salt Lake City! Our team at Think Circuits Engineers is working on a system to autonomously detect and suppress incipient wildfires in 10 minutes or less across thousands of square kilometers. Looking forward to connecting with investors, stakeholders, and other participants to share our vision and progress. To find out more information about our plans please visit tcwildfire.com, or follow us at Think Circuits Engineers!

We're #hiring a new Senior Electrical Engineer in Oakland, California. Apply today or share this post with your network.