01 / Mobile
The field companion.
Cross-platform label capture, tactile feedback, and cloud-connected inspection results in a native mobile experience.
React NativeExpo
Full-stack development & AI vision
I build the whole picture.
Hey, I’m Sai Harsha. I connect thoughtful interfaces, cloud systems, and machine intelligence. From the first pixel to the last API call.
INTERFACES → INTELLIGENCE
THREE LAYERS. ONE CONNECTED EXPERIENCE.
01 / Selected work
One ecosystem.
Every layer.
The interesting part isn’t just building an app. It’s making the entire system work together.
Explore the sourceLegal Metrology Vision System
From a packaging photograph to a connected inspection workflow.
FIELD / MOBILE
Label scanner
SAMPLE PACKAGEEveryday
essentials. NET QTY. 250 g
essentials. NET QTY. 250 g
Position the label inside the frame
From the field.CAMERA → CLOUD
01 / The human interface
Start where
Start where
the work happens.
A mobile field-scanner brings label capture to the inspection site. Native camera access and gallery uploads connect the physical package to the vision service.
React NativeExpoTypeScript
PROCESS / INTELLIGENCE
IMAGE IN. STRUCTURE OUT.
vision → structured data
// simplified example
{
"label":
"sample package",
"declarations": [
"net quantity",
"manufacturer"
],
"regions":
"2D bounding boxes" }
02 / The intelligence layer
Give unstructured
Give unstructured
images structure.
A Python vision service uses Gemini to extract packaging declarations and locate them with bounding boxes, returning structured results for the client applications.
PythonFastAPIGemini
REVIEW / WEB
LM / Command center
The bigger picture.
SAMPLE INSPECTION · NOT LIVE DATA
Net quantityEXTRACTED
ManufacturerEXTRACTED
Label declarationsFOR REVIEW
Scan history · User roles · Reports
03 / The connected workspace
Turn individual scans
Turn individual scans
into a shared view.
A Next.js dashboard brings inspection history, role-based access, and reporting together. Prisma and PostgreSQL provide the persistence layer behind the web experience.
Next.jsPrismaPostgreSQL
Built across the stack.
ONE PROJECT / THREE COMPONENTS
02 / AI vision
The vision engine.
A focused microservice that turns packaging photographs into extracted declarations and spatial coordinates.
FastAPIGoogle Gemini
03 / Web & cloud
The command center.
A connected workspace for reviewing scans, managing user access, and keeping inspection records in view.
Next.jsPostgreSQL
02 / The toolkit
Different tools.
Same curiosity.
The technologies behind my public projects, grouped by what they help me build.
01
Interfaces
TypeScript · React · Next.js
React Native · Expo ·
Tailwind CSS
02
Intelligence
Python · FastAPI · Gemini
Multimodal vision · Pillow
03
Data & access
PostgreSQL · Prisma · Neon
NextAuth.js · Role-based
access
04
Delivery
Vercel · Render · Expo EAS
Git · GitHub · OTA
updates
A little more human.
A lot more curious.
A lot more curious.
Allika Venkata Sai HarshaDEVELOPER / BUILDER / STUDENT
03 / Beyond the code
Curious by nature.
A builder by choice.
I’m a B.Tech Computer Science & Engineering student at GITAM University, based in Hyderabad. My focus is full-stack cloud ecosystems and AI vision systems.
I’m drawn to the space between an interface and the infrastructure behind it—where a useful idea becomes a connected experience.
04 / What’s next?
GOOD IDEAS START WITH A CONNECTION.
Explore the code, follow what I’m building, or bring a fresh perspective to a project.
https://github.com/AllikaSaiHarsha