Education · 2021 — 2025
9.43/10 CGPA
Pandit Deendayal Energy University
Bachelor of Technology, Computer Science & Engineering
Software Engineer — Backend · Cloud · AI/ML
Khushi Desai — Software Engineer, Backend, Cloud and AI/ML
I build scalable backends, cloud platforms, real-time systems, secure services and AI-powered products.
↳ ASP.NET Core · Node.js · FastAPI · REST · microservices
Software engineer who owns systems end to end: service and API design, event-driven pipelines, databases, identity and security, cloud delivery and observability, plus AI/ML where the product needs it. Comfortable in any domain; so far healthcare, IoT, fintech and AI.
- ASP.NET Core
- Node.js
- FastAPI
- Express.js
- Redis Streams
- MQTT
- Socket.IO
- PostgreSQL
- SurrealDB
- MongoDB
- Docker
- Terraform
- Jenkins
- AWS
- Traefik
- Keycloak
- Kong Gateway
- Prometheus
- Grafana
- WebRTC
- PyTorch
- Faster-Whisper
- YOLO
- LLMs
01About — end to end
About Khushi
Khushi DesaiAhmedabad, IN
Good software doesn't stop at the endpoint. I build the whole path — the API, the queue, the database, the cloud it runs on and the intelligence on top — and make it hold up in production.
I'm a software engineer at Glide Technology. Most of my work is backend and cloud: designing services and APIs, event-driven pipelines and data models, wiring in identity and security, and shipping it all with containers, infrastructure-as-code and monitoring. When a product needs intelligence, I integrate the AI/ML services too. I've worked across healthcare, IoT, fintech and AI, and I pick up new domains quickly.
- Shipping production software at Glide Technology
- 1+yrShipping production software at Glide Technology
- Industry platforms across AI, IoT, healthcare & fintech
- 4Industry platforms across AI, IoT, healthcare & fintech
- B.Tech CSE CGPA, Pandit Deendayal Energy University
- 9.43/10B.Tech CSE CGPA, Pandit Deendayal Energy University
- Certified Developer – Associate, 2026
- AWSCertified Developer – Associate, 2026
02Work — select a project
Selected work,told as systems.
Backend, cloud, security and AI across healthcare, IoT, fintech and AI products. Pick a project for the overview; each featured case opens into a full interactive case study.
Featured case studies
Real-time audio intelligence & AI conversation insights
A distributed .NET + Python backend that turns live conversations into speaker-attributed transcripts, LLM insights and long-lived session memory.
Backend & cloud
- ASP.NET Core 10 orchestration: sessions, ingestion, persistence, LiveKit tokens
- Redis Streams consumer groups & job queues for async, decoupled processing
- FastAPI service boundary between .NET and Python ML workloads
- SurrealDB persistence; Docker Compose + Traefik load-balanced STT replicas
AI / ML
- Faster-Whisper automatic speech recognition
- NVIDIA NeMo speaker diarization
- PyTorch + Torchaudio audio processing services
- Local Mistral GGUF LLM for summaries, sentiment, categories & action items
- C#
- ASP.NET Core 10
- Python
- FastAPI
- Faster-Whisper
- NVIDIA NeMo
- PyTorch
- Torchaudio
- +8 more
Inside: Live pipeline replay · interactive architecture · 8 contributions · 4 engineering decisions
More builds
Case 01 / AI platform · distributed backendGlide Technology · Jul 2025 — Present
ABOVE AI
Real-time audio intelligence & AI conversation insights
A distributed .NET + Python backend that turns live conversations into speaker-attributed transcripts, LLM insights and long-lived session memory.
Backend & cloud
- ASP.NET Core 10 orchestration: sessions, ingestion, persistence, LiveKit tokens
- Redis Streams consumer groups & job queues for async, decoupled processing
- FastAPI service boundary between .NET and Python ML workloads
- SurrealDB persistence; Docker Compose + Traefik load-balanced STT replicas
AI / ML
- Faster-Whisper automatic speech recognition
- NVIDIA NeMo speaker diarization
- PyTorch + Torchaudio audio processing services
- Local Mistral GGUF LLM for summaries, sentiment, categories & action items
- C#
- ASP.NET Core 10
- Python
- FastAPI
- Faster-Whisper
- NVIDIA NeMo
- PyTorch
- Torchaudio
- Mistral GGUF
- LiveKit
- WebRTC
- Redis Streams
- +4 more
Inside: Live pipeline replay · interactive architecture · 8 contributions · 4 engineering decisions
03Skills — stack & industry fit
No skill bars.Just where each tool shipped.
55 tools organised by system layer, plus how my work maps onto where the industry is heading in 2026.
Evidence
Select a tool on the map. Tools marked with ● have shipped in a project on this site.
Tap a tool to see where it shipped. ● = used in a project on this site.
L1Services
APIs & backend runtimes
L2Real-time
Streams, events & media
L3Data
Persistence & state
L4Cloud & platform
Infrastructure & delivery
L5Trust & observability
Identity, security & telemetry
L6Intelligence
Models, inference & ML pipelines
L7Interface
Real-time UIs & visualisation
L8Languages
Daily drivers first
04Background — experience, education & credentials
From pixelsto production systems.
Design taught me to care about the person on the other end. Engineering taught me to make it hold up in production. Each role went deeper into the stack.
Experience
Scroll to move through the timeline →Swipe through the timeline →
Education & credentials
Certified · Sep 2026
AWS Certified Developer – Associate
Amazon Web Services
Verify on Credly ↗ (opens in a new tab)Certifications
Recognition · Nov 2023
New India Vibrant Hackathon — Finalist
Designed a Medicinal Plants Awareness app: identify a plant from an uploaded image and surface its medicinal information.
05Contact — open channel
Have a systemthat needs building?
I'm looking for software engineering roles in backend, cloud and AI/ML, in any domain. APIs, distributed services, infrastructure or AI features: if it has to work in production, I'm interested. The fastest way to reach me is email.