Backend & APIs
FastAPI services with layered API → service → repository design, JWT auth and RBAC.
☀️ Good Morning Welcome to my digital workspace.
Yash Gaidhane
I work as a Full Stack Developer
Full Stack Developer focused on Python/FastAPI, React, Flutter, APIs, AI-assisted systems and production deployment. I take features from schema and backend logic through UI, testing and deployment.
Open to opportunities
Full Stack Developer · Backend Developer · Software Engineer
PythonFastAPIReact
TypeScriptFlutterPostgreSQL
RedisDockerGit
Scroll to explore
About · developer profile
Building reliable software from idea to deployment.
I design full-stack applications, backend systems, AI-assisted platforms, and cross-platform products. This content is physically pulled upward by the rocket as the visitor scrolls.
Open to workFull Stack · BackendCurrently: Python / FastAPI
Workspace sequence · live
Engineering ideas into working systems.
A short interactive sequence introduces the projects, tools, and problem-solving mindset behind this portfolio.
workspace/init.sh
$ initializing developer workspace...
$ connecting BlackBox crash-diagnosis pipeline
$ loading Suraksha safety workflows
$ starting support-ticket automation
$ verifying APIs, data flows, and deployment
✓ Workspace ready — continue exploring.
About · who I am
I build software that
holds up.
I’m a Full Stack Developer focused on backend systems, APIs, AI-assisted workflows and cross-platform applications. I enjoy the unglamorous engineering: schemas, queues, authentication, migrations, testing and deployment.
When a system isn't sure, I'd rather it say so than invent an answer.
AI & RAG systems
Retrieval-augmented diagnosis and LLM triage with human handoff below confidence thresholds.
Cross-platform
Flutter for Android and iOS, including background services and wearable companions.
Data & queues
PostgreSQL schemas and migrations, Redis and Celery for asynchronous processing.
Cloud & DevOps
Dockerised services, GitHub Actions pipelines and pytest suites.
Embedded systems
ELF/DWARF symbolization for STM32 and FreeRTOS crash reports using pyelftools.
By the numbers
0Tests in CI
0Test coverage
0Featured projects
0Languages I ship in
Featured project · BlackBox
Hex dump in.
Root cause out.
BlackBox ingests STM32/FreeRTOS crash reports, resolves addresses through ELF/DWARF symbolization and produces AI-assisted diagnosis — or returns uncertain when evidence is insufficient.
BlackBox
AI-powered embedded crash diagnosis platform
- Problem
- Firmware crashes arrive as raw hex stack dumps. Matching addresses against the correct build's symbols is slow and error-prone.
- Solution
- Symbolizes crash traces, groups recurring signatures and uses a retrieval-augmented LLM pipeline for diagnosis. Unsupported cases return “uncertain” rather than fabricate an answer.
- AI-assisted diagnosis
- ELF/DWARF symbolization
- Crash signature grouping
- JWT authentication + RBAC
- Celery / Redis async workers
- Layered API → service → repo
- 400+ tests at 82% coverage
BLACKBOX // INTERACTIVE DIAGNOSIS
crash_report.raw
diagnosis
Press Diagnose.
Canned data — the real pipeline runs on a server.
Selected work · 4 projects
What I build.
A mix of backend, AI, mobile and systems work. The goal is always the same: solve a real problem and make the system explainable.
CRASHHardFault
0x0800a3f4SYMBOLvSensorRead
bmp280.c:214AI VERDICTStack overflow detectedconfidence: 94%
BlackBox
AI-powered crash diagnosis● Live · Open Source
Turns raw embedded crash dumps into symbolized stack traces and AI-assisted root-cause diagnoses with async processing and confidence-aware output.
Impact: 400+ automated tests · 82% coverage · production deployment · confidence-aware diagnosis.
SURAKSHA
SOS
● GPS LIVE
Trusted contacts 3
Suraksha
Real-time personal safety platform● Android v1.0.0 · Live Guardian Tracker · Open Source
Flutter + Firebase safety platform with SOS alerts, shake detection, trusted contacts, real-time GPS sharing, journey monitoring, community safety reporting, history and privacy controls.
Impact: Release APK tested on a physical Android device · real GPS movement tracking · automatic journey escalation · browser-based guardian tracking · Demo Mode authentication.
SOS + ShakeCountdown, emergency escalation and safety resolution.
Live LocationReal-time GPS tracking with guardian preview.
Safety ToolsJourney monitoring, contacts, history and community reports.
SUPPORT OPSLIVE
#4182Payment failureHIGH
#4179Login issueMED
#4172Refund requestLOW
Support Ticket System
Auto-triage helpdesk API● Open Source · Backend Project
Full ticket lifecycle with roles, immutable event logs, rule/LLM triage, SLA policies, CSAT reporting and PII redaction.
Impact: RBAC ticket lifecycle · SLA escalation · rule/LLM triage · PII-aware processing.
Ticket QueuePriority and ownership overview.
Auto-TriageRules and LLM confidence routing.
SLA PanelEscalation and breach tracking.
WAR CREEDUNREAL ENGINE 5 · ENVIRONMENT
War Creed
3D game environmentEnvironment and level design on a team project in Unreal Engine 5, including terrain, vegetation, materials and lighting.
GitHub profile ↗BlackBox case study · deep dive
Crash reports are not explanations.
Input
Embedded crash reports often contain fault registers, task information and raw addresses. Without the matching firmware build, those addresses are difficult to interpret.
Goal
Resolve addresses to source symbols, identify repeatable crash signatures and generate a diagnosis only when the available evidence supports it.
Architecture
From device to diagnosis.
STM32 / FreeRTOS
→
FastAPI
→
Celery + Redis
→
PostgreSQL
ELF / MAP upload
→
pyelftools
→
RAG retrieval
→
LLM diagnosis
API flow
Diagnosis pipeline.
1
Register build
Upload the matching ELF or MAP artifact with firmware and build metadata.
2
Ingest crash
Receive device, firmware, fault, register and stack-address data through the API.
3
Symbolize
Resolve addresses to function names and source lines using DWARF information.
4
Group
Create a stable signature to connect recurring crashes across devices and builds.
5
Diagnose
Retrieve relevant knowledge and return a root cause, recommended fix and confidence score—or “uncertain” when evidence is weak.
Example
What the output looks like.
diagnosis.json
{
"root_cause": "NULL pointer dereference in SensorTask",
"summary": "SensorTask crashed due to an invalid pointer access.",
"recommended_fix": "Review pointer validation, task stack sizing and buffer handling.",
"confidence_score": 0.94
}Engineering decisions
Layered backend
API → service → repository separation keeps HTTP handling, business logic and persistence independently testable.
Asynchronous processing
Celery and Redis move symbolization and diagnosis work away from request-response latency.
Confidence-aware AI
Unsupported cases return “uncertain” instead of fabricating a root cause.
Security
JWT authentication, role-based access control and scoped device API keys protect administrative and ingestion workflows.
GitHub · activity
Building in public.
Live public activity from my GitHub profile. The contribution image loads from GitHub when the visitor is online.
The graph is fetched from a third-party contribution-chart service and may be unavailable offline.
Location
Raipur, India
Available for remote opportunities and collaborative software projects.
Toolkit
What I work with.
Tools used to ship something real — no self-assessed percentage bars.
- Languages
- PythonTypeScriptDartC++SQL
- Backend & data
- FastAPISQLAlchemyREST APIsCeleryJWT / RBACPostgreSQLRedisFirebase
- Frontend & mobile
- ReactFlutterTailwind CSS
- AI & systems
- RAG pipelinesLLM integrationpyelftoolsELF / DWARFSTM32 / FreeRTOS
- DevOps & tooling
- DockerGitGitHub ActionspytestLinux
Right now · current focus
What I’m building on next.
I’m strengthening the areas that matter most for production-ready full-stack engineering.
Designing cleaner service boundaries, scalable APIs, background processing and maintainable application architecture.
Improving containerized deployments, CI/CD workflows, production reliability and observability.
Building practical RAG and LLM-assisted features with confidence checks, retrieval quality and safe fallbacks.
Recruiter snapshot · why hire me
I don’t stop at “it works.”
I focus on turning requirements into complete, testable and deployable software rather than isolated code samples.
End-to-end ownership
I can work across database design, APIs, business logic, frontend/mobile interfaces, testing and deployment.
Backend-first problem solving
My strongest work includes FastAPI systems, authentication, async workers, PostgreSQL, Redis and API-driven product architecture.
Projects with real engineering depth
BlackBox, Suraksha and the Support Ticket System demonstrate AI integration, mobile workflows, background processing, security and production-minded design.
Let’s connect
Looking for a developer?
Let’s talk.
Open to Full Stack Developer, Backend Developer and Software Engineer opportunities.
Have a role where I can build APIs, full-stack products or AI-assisted systems? I’d be happy to discuss it.
yashgaidhane2@gmail.com