Shiphura Arinda
System Administrator
Node JS
PHP
JavaScript
Python
Active 7 days ago
+256 (768) 027886 Invite to interview
Message
Download CVCV
Overview
Technical skills
Timeline
Roles
Overview
Backend API and ML engineer (mid-level) focused on building end-to-end ML-backed screening services with a practical strength in prototyping robust data ingestion and model inference pipelines. The strongest proven skill is production-oriented ML inference and reporting driven by app/ml/train.py and the inference endpoints in app/main.py which include model selection, metrics reporting, and batch upload parsing. Public code shows limited operational hardening - there is no evidence of authentication/authorization, migration histories, advanced scaling, or observability beyond basic logging and saved metrics files.
Technical skills
Node JS• Middle
PHP
JavaScript• Junior
Python• Middle
Node JS
Express
Mongoose
BullMQ
PHP
WooCommerce
WordPress
Python
FastAPI
SQLModel
Pydantic
Databases
MySQL
MongoDB
Frontend
Bootstrap
React.js
Tailwind CSS
PostCSS
Vite
React Router
React Hook Form
Zod
Recharts
autoprefixer
DevOps
Docker
Git
QA
Postman
Mocha
Analytics
Matplotlib
AI/ML
NumPy
Pandas
Scikit-learn
Timeline
System Administrator
•
Middle
Nunu Fund Uganda
•
Full-Time
Provided first-level IT support for staff, resolving hardware, software, access, and application issues. Managed user accounts, roles, permissions, and supported Google Workspace access and email configuration. Maintained and monitored internal systems, performing updates, backups, and security checks while troubleshooting system errors and coordinating with developers on improvements. Created technical documentation, user guides, and operational reports to support adoption and decision-making.
Google Workspace
California Institute of Technology (Caltech)
Other •
Science in softwae Engineerinf
Intern Developer / IT Support Assistant
•
Junior
Outdoorsy Uganda
•
Full-Time
Supported website users and business applications by troubleshooting errors, performance issues, and plugin conflicts. Maintained a WordPress site, handling plugins, themes, integrations, updates, backups, and security improvements. Supported e-commerce operations using WooCommerce and documented technical issues and implemented fixes.
WordPress
WooCommerce
Full-Stack Developer
•
Middle
Happy Hoe Grocery Store
•
Full-Time
Built a stock management system to help inventory tracking and daily record keeping. Implemented core features for adding, editing, searching, and managing inventory records through a user-friendly interface for business users. Supported improved business operations via digital management of inventory data.
Middle Backend Developer
Confidence: Medium API Engineer
Backend API and ML engineer (mid-level) focused on building end-to-end ML-backed screening services with a practical strength in prototyping robust data ingestion and model inference pipelines. The strongest proven skill is production-oriented ML inference and reporting driven by app/ml/train.py and the inference endpoints in app/main.py which include model selection, metrics reporting, and batch upload parsing. Public code shows limited operational hardening - there is no evidence of authentication/authorization, migration histories, advanced scaling, or observability beyond basic logging and saved metrics files.
API Design
3/10
How well APIs are designed
Reasonable REST API design with input schemas and predictable error handling, but no versioning, idempotency keys, or advanced API lifecycle patterns.
Evidence
AI-powered-early-Autism-screening-system/backend/app/main.py: /predict, /screenings, /screenings/batch endpoints with HTTPException usage
AI-powered-early-Autism-screening-system/backend/app/schemas.py: Pydantic-based request/response models
AI-powered-early-Autism-screening-system/frontend/src/lib/api.ts: client-side API wrapper and typings
Data Layer & Database
3/10
Working with databases
Clear data model usage and basic transactional flows using SQLModel/Session and Mongoose, but no migration history, explicit transaction management, or hand-tuned SQL.
Evidence
AI-powered-early-Autism-screening-system/backend/app/models.py: Screening SQLModel model
AI-powered-early-Autism-screening-system/backend/app/db.py: SQLModel Session and engine usage (get_session/init_db)
trackingWorkers--profiles/models/Profile.js and config/db.js: Mongoose schema and connection logic
Scalability & Performance
2/10
Handling load and speed
Some attention to model training performance and use of efficient ML libraries, but the runtime service lacks caching, async/queue-based batch processing, and explicit scaling or rate-limiting strategies.
Evidence
AI-powered-early-Autism-screening-system/backend/app/ml/train.py: model training pipelines and model selection saved to disk
AI-powered-early-Autism-screening-system/backend/app/main.py: /screenings/batch processes rows sequentially via pandas.iterrows
System Architecture
3/10
Overall system structure
Reasonable module separation for ML, API, schemas and frontend, showing deliberate structure for an end-to-end prototype, but no evidence of multi-service orchestration, deployment topology, or advanced decomposition trade-offs.
Evidence
AI-powered-early-Autism-screening-system/backend/app/: clear separation into ml, models, schemas, db and main
AI-powered-early-Autism-screening-system/frontend/src/lib/api.ts: typed client and contract alignment with backend
Security & Auth
3/10
Protecting data and access
Basic input validation and safe parsing for uploaded files and model inputs, but no authentication/authorization, secrets management, or hardened defenses against advanced threats.
Evidence
AI-powered-early-Autism-screening-system/backend/app/main.py: file type checks and robust CSV/ARFF parsing with validation and HTTPException for bad inputs
AI-powered-early-Autism-screening-system/backend/app/schemas.py: pydantic schemas for ScreeningInput and PredictionOutput
trackingWorkers--profiles/routes/profile.js: multer fileFilter to allow only images (upload hygiene)
Reliability & Observability
2/10
Stability and monitoring
Basic try/except error handling and status reporting are present, but there is minimal observability, no retry/backoff strategies, no structured correlation ids, and no graceful shutdown patterns.
Evidence
AI-powered-early-Autism-screening-system/backend/app/main.py: try/except blocks around file parse and model metric loading
AI-powered-early-Autism-screening-system/backend/app/ml/train.py: metrics and report generation written to disk for dashboard use
trackingWorkers--profiles/config/db.js: process.exit on DB connect failure (basic failure handling)
Expertise
Backend AI & LLM• Middle
Python• Middle
Node.js• Middle
Industries
Data & Analytics• Middle
Health Care• Middle
Technologies
Python• Middle
PHP
Node JS• Middle
MySQL
MongoDB
Express
WordPress
FastAPI
WooCommerce
Git
Docker
Pydantic
Mongoose
BullMQ
SQLModel
Recommendations
- Develop ML-backed REST APIs and data ingestion pipelines where you own model training, artifact management, and inference endpoints (extend current FastAPI code).
- Build prototype healthcare screening products or analytics dashboards that require careful data parsing, label handling, and lightweight explainability (use existing dataset loaders, features, and report generation).
- Implement production hardening for services: add structured logging/correlation ids, timeouts, retries with backoff, and a simple deployment/migration plan (CI/CD, DB migrations).
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories:
Middle Frontend Developer
Confidence: Medium App Engineer
Frontend app engineer (mid-level) focused on building React single-page applications with typed client-side API integration. The strongest proven skill is typed API surface design and client-server contracts as implemented with zod schemas and a typed http client in frontend/src/lib/api.ts. There is limited evidence of automated testing, advanced state machines, deliberate performance measurement, or production-grade observability in the public code.
UI Component Architecture
2/10
How interface parts are built
Minimal custom component architecture visible; a few utility modules suggest some structure but no evidence of a deliberate component library, composition patterns or complex lifecycle hooks.
Evidence
AI-powered-early-Autism-screening-system/frontend/src/lib/format.ts
Arinda-web/assets/js/main.js
Responsive & Cross-browser
4/10
Works on all screens and browsers
Responsive layouts and mobile nav behavior are implemented with media queries and JS toggles; Bootstrap-based pages and Tailwind base show practical responsive work but not advanced feature detection or RTL/i18n readiness.
Evidence
Arinda-web/assets/css/style.css
Arinda-web/assets/js/main.js
Performance Optimization
2/10
Speed of the interface
Little explicit performance engineering or measurement; standard modern build tools (Vite, Tailwind) are used but there is no evidence of bundle analysis, code-splitting strategy or measured LCP/INP optimizations.
Evidence
AI-powered-early-Autism-screening-system/frontend/vite.config.ts
AI-powered-early-Autism-screening-system/frontend/src/styles.css
Accessibility & Semantics
3/10
Usable for everyone
Some semantic and a11y-friendly marks exist (role=progressbar, labelled form controls, alt text), but no evidence of ARIA on custom widgets, focus-management patterns or automated a11y checks in CI.
Evidence
Arinda-web/index.html (progress bars with role="progressbar")
personal-portfolio/index.html (form fields with labels and placeholders)
State Management & Data Flow
3/10
Managing data in the app
Clear, typed client-server API surface with validation schema and typed responses; good server-state typing and basic error handling but no request cancellation, cache invalidation, optimistic updates or state machines.
Evidence
AI-powered-early-Autism-screening-system/frontend/src/lib/api.ts
AI-powered-early-Autism-screening-system/frontend/src/lib/api.ts (ScreeningInputSchema using zod)
UX & Visual Polish
3/10
Look and feel quality
Visually polished pages and interactive behaviors are present, but much of the visual polish derives from templates and libraries; UX edge states and experiment-driven perceived performance improvements are not evident.
Evidence
Arinda-web/assets/css/style.css
Arinda-web/assets/js/main.js
Expertise
Frontend Architecture & Build Tools• Middle
Industries
Health Care• Middle
Technologies
Tailwind CSS
Bootstrap
React.js
Vite
PostCSS
Zod
Recharts
React Hook Form
React Router
autoprefixer
Recommendations
- Lead frontend feature work for React-based SPAs that integrate with REST/ML backends, owning client API contracts and form flows.
- Implement and harden the screening UI - add request cancellation, race handling, optimistic UI with rollback and explicit loading/error/empty states.
- Build a small component library or design-system tokens for reuse and document component boundaries and accessibility behavior.
- Add automated tests (unit and integration), lightweight performance audits (Lighthouse) and simple CI checks for a11y and linting.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories:
Junior QA Engineer
Confidence: Medium Generalist
A Junior generalist developer focused on basic unit testing and front-end tooling. The strongest proven skill is writing simple unit tests with Mocha and Chai as shown in arindashiphura/tdd/test/test.js and arindashiphura/tdd/test/multiply.js. There is no public evidence of integration testing, CI pipelines, performance/load testing or sophisticated negative-path test design.
Test Automation Frameworks
2/10
Building automated tests
Basic unit test automation using Mocha and Chai is present but there are no fixtures, factories, parallelization, selector discipline or E2E/browser automation patterns; evidence is limited to straightforward function tests.
Test Coverage & Strategy
1/10
What and how to test
Test coverage is limited to small happy-path unit tests without explicit risk-based structure, negative-paths, boundary tests or property-based checks.
Evidence
API & Integration Testing
Testing how parts work together
Not evidenced in public code
Performance & Load Testing
Testing speed under load
Not evidenced in public code
Bug Reporting & Analysis
Finding and describing bugs
Not evidenced in public code
CI Test Integration
1/10
Running tests automatically
NPM scripts include a test command which supports local test runs, but there is no CI workflow, matrix, per-test retry policy, artifact capture or integration with CI systems.
Evidence
Expertise
Unit & Component Testing• Junior
Technologies
JavaScript• Junior
Postman
Mocha
Recommendations
- Build and maintain small front-end applications with unit tests and linting to ensure quality in UI components.
- Expand unit tests to include negative and boundary cases and add simple property-based or invariants for core functions.
- Add basic CI integration with per-test reporting and occasional artifact capture to turn local scripts into reproducible pipelines.
- Practice test infrastructure patterns such as fixtures, isolated data setup/teardown and simple mock servers for API interactions.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories: