Invitation to lead accuracy at HyperTrack
Ground truth
So, Johnny Tucker logs in to his shift at 8:30AM while at home. He drives 10 minutes and 2.5 miles to the office. He walks 128 steps to the parking lot to pick up the service vehicle and start visiting customers. In the process of doing his three visits, he goes through four tunnels, climbs five floors, walks 685 steps, drives on roads with tall buildings on the side, breaks the speed limit twice, loses network connection thrice and gets delayed by 8 minutes for the last visit. Johnny breaks for lunch at 12:30 at his favorite eatery near the office. Tracking stops.
Ground reality
The team that built Johnny’s work app counts on HyperTrack. To gather this ground truth and break the data into relevant parts. For his supervisor and the three customers to consume. At the time when it matters most to them. Business depends on it.
If it works 85% of the time, that means it is wrong 1 out of 6 times. My team is not going to use it. I am not going to bet my business process on it.
Think of it as voice recognition in 2005. We all knew it was important. But it was not good enough for us to talk to in our living rooms.
The accuracy challenge
Deliver accuracy that businesses can rely on.
HyperTrack has access to a variety of APIs in the device OS to get relevant data about movement. We need to figure out what to use when. Over-using them will kill battery. Under-using will miss the ground truth.
Data that gets to the cloud is inherently noisy and unreliable. Making it contiguous and applying laws of physics will clean out the noise. Over-correcting it will amplify the noise. Under-correcting will leave it noisy.
Location, sensor, maps and customer’s application data must be merged to infer places visited, routes taken, parking, delays, unexpected behavior, speeding, entries/exits, and nearly all parts of the story that applications and users care for. Serializing will add latency and miss loopbacks. Parallelizing will bloat data and resource costs.
The human challenge
- Problem solver: You love simple open-ended problems with randomness and probabilities, that cannot be solved with deterministic systems. You think first principles, often laterally. You feel the compulsion to solve the current Jane Street puzzle except you want to work in the SF Bay Area. You apply calculus, combinatorics, algebra and geometry to your favorite recreational activities.
- Adventurer: You consider a pre-inflection-point startup a life experience. You gravitate towards picking up and solving one of the top three problems around you and thrive in the pressure that comes with it. Your goal is to produce magic that makes people wonder how in the world you did that. You expect to fail majority of the times you try to do so.
- Clear communicator: You understand that ideas are only as good as others’ ability to get it. Great ideas take a village to deliver, and great communication is a pre-requisite to rally the team behind them. Clarity of thought is the strength that lets you and others dig deeper to find answers. Driving outcomes is more important to you than giving it your best shot.
Score yourself from 1 to 5 on each of the above. Drop everything and call me if you scored 10+.
Other pre-requisites
- Be a culture fit
- Be a good developer
See this for more.
Basics about us
- WeWork at 353 Sacramento in SF FiDi, on-site 5 days
- 7 in SF, 5 in Ukraine, 3 in transit from Bangalore to SF
- Health, dental, vision, HSA, accrued leaves, coffee/beer, off-sites, puns
- We will relocate you, especially if you are authorized to work in the US
- Millions of movement data pings a day, growing fast
- Users are developers, building magic for their internal/external users
- We track commerce, do not track consumers, respect privacy
- More here
Last thing…
We have found a deep pain that lots of customers worldwide will pay us to solve. Our superpowers are infrastructure, data, mobile, design, and developer experience. Adding accuracy as a superpower will complete us.
kashyap@hypertrack.com
Hit me!
