Gather AI
YEAR 2021
I joined Gather AI, a world leader in drone analytics using AI, just before the pandemic. It was a major shift from a company with hundreds of thousands of employees to a startup with fewer than ten. In my role as Product Architect and HCI Architect, I was responsible for each new feature from its very inception to its completion. This end-to-end cycle—from ideation and research to development, testing, and customer engagement—got me fired up for the challenge.
ROLE :
Product Architect, HCI Researcher
TEAM :
Worked with the CEO, CTO, and Chief AI & Autonomy Scientist
KEY TECHNOLOGIES
Autonomous Drones
3D AI Analytics
Unified Dashboard
Spatial Mapping
TOOLS & SOFTWARES USED:
Capturing Reality
DJI SDK
SFM Algorithms
CloudCompare
Adobe Creative Suite
“Leading the development of drone AI analytics for 3D worlds.”
Methodology & Approach
Until then, Gather AI had used its drones for 2D analytics in indoor environments. My mission was to extend their reach into the third dimension with spatial, 3D AI analytics using autonomous flying drones. My previous work in augmented and virtual reality gave me a head start, so I sought to build on that knowledge. I started by implementing their existing 2D analytics framework on the ground to gain a thorough understanding of the end-user's experience, bridging the gap between vision and execution.
DEPLOYMENT - Actual Flight Environment 🠈
-Autonomy
“Flying a drone autonomously indoors requires surgical accuracy.”
OVERVIEW - Components and systems that work together to make up the Product 🠉
Research & Prototype
In the initial research phase, I tested various techniques to extract a volumetric map using an algorithm called SFM (Structure From Motion). Once I had a working prototype, I collaborated with the Autonomy Scientist and CTO to strategize a flight path for the drone to autonomously collect visual data. I built a small yet powerful server to process this data, perform volumetric scene reconstruction, and upload the information to a web dashboard where the end-user could interact with it and extract insights. The entire workflow was made functional so that a user could simply select a location on a map and run a flight in real-time or schedule it for a later date.
FLIGHT PATH - A reconstructed 3D model created from an autonomous orbital flight path 🠉
-Research
“The blueprint for our features was drawn from the very people who would be using them.”
INTERVIEW - Interview sessions were conducted via video calls 🠉
I conducted detailed interviews and created questionnaires based on existing workflows, pain points, and expectations of various project partners as well as clients, from a groundworker to the Director of Operations.
These interviews were instrumental in gathering key insights, which I then documented and mapped into a feature list.
ARCHITECTURE - The flow of data within interconnected modules and their collaborative function.🠉
DASHBOARD - The end-user interface through which users interact with the product 🠉
Final Outcome
Once a functional workflow was ready, the next task was to optimize it by quantifying metrics at each stage of the pipeline and identifying KPIs (Key Performance Indicators) and KEIs (Key Experience Indicators). In the end, we had a completely new feature built from the ground up that garnered significant client interest.
Later, I took charge of creating a promotional video, a crucial step for documenting the process and giving it life as a structured narrative. The film was instrumental in securing a contract with a client and helped the company garner more funding during a tough time.
PROJECT LIVE - The Project is being used in a live scenario 🠉
FINAL CUT - A detailed video that covers the product's features, use cases, and customer testimonials 🠋
