AWS-CERTIFIED SOFTWARE DEVELOPER

I turn businessneedsintopractical software.

I design customer-facing software and cloud workflows using AWS, Nuxt, TypeScript, and Python. My recent work covers secure serverless APIs, document systems, and grounded generative AI.

Explore selected projects AWS Certified AI Practitioner · Kozhikode, India

Kozhikode, India
11.2588° N / 75.7804° E

( 01 / About )

I turn business needs into clear, useful software—from customer-facing interfaces to AWS and generative AI workflows.

I’m Riswin Mohammed, a software developer based in Kozhikode. My work spans a flooring retail website, a supplier-invoice application, and a company knowledge assistant. On Floorsir, my responsibilities included client acquisition, UX direction, frontend development, and iteration.

I’m particularly interested in building secure, customer-facing applications on AWS. My recent work explores authentication, serverless APIs, document workflows, retrieval-augmented generation, and human-controlled AI systems—while keeping the business problem central to the architecture.

AWS Certified AI Practitioner certificate issued by Amazon Web Services

View certificate

Certification

Amazon Web Services

AWS Certified AI Practitioner

Issued: 30 Aug 2026
Valid through: 30 Aug 2029

ID: 408c95adc9bc486a813bb3b269c3ee5a

Verify with AWS

( 02 / Selected work )

Built to
matter.

Recent experiments
& products / 2023—26

Primary AWS case study

AI-powered company knowledge assistant2026Public demo · Sign-in required

Contexta

Helping employees find answers across internal company documents. Contexta combines a Nuxt interface with Amazon Bedrock retrieval and generation, with citations that link answers back to their sources.

The public demo uses fictional company documents, citations, and a five-question daily limit. The described implementation includes Cognito authentication, protected APIs, permission-aware retrieval, document ingestion, conversation history, feedback, guardrails, monitoring, and RAG evaluation.

Technical decision

Separate ingestion from question answering: documents are retrieved through a Bedrock Knowledge Base before the model returns a cited answer, while demo authentication and usage limits protect access.

Open demo

AWS architecture

Two separated paths keep document ingestion distinct from authenticated question answering.

Described implementation

Question path

Nuxt clientCognito sessionAPI GatewayLambdaBedrock Knowledge BaseCited answer

Document path

Private S3Document ingestionBedrock Knowledge BasePermission-aware retrieval

My role

System architecture
Full-stack development
AWS integration
RAG evaluation

Technology

NuxtVueTypeScriptPythonAWS LambdaAPI GatewayAmazon BedrockBedrock Knowledge BasesS3DynamoDBCognitoIAMCloudWatch

Featured project

Customer-facing flooring retail2024

Floorsir

I translated a flooring retailer’s customer journey into a responsive catalogue experience covering carpets, parquet, vinyl, raised flooring, sports flooring, services, and accessories.

The visible flow supports category browsing, product exploration, designer sign-up, home or store visits, sample requests, and WhatsApp enquiries. The public site does not establish a transactional checkout flow.

Technical decision

Treat the catalogue and enquiry journey as the product surface, avoiding a checkout claim the public experience does not support.

View live site

My role

Client acquisition & sales
UX direction
UI design
Frontend development
Responsive implementation
Iteration

Technology

VueNuxtTypeScriptTailwind CSSGSAP

Featured project

Invoice management for small businesses2026Live demo · AWS workflow unverified

Invoxa

A working supplier-invoice workflow with upload and preview, manual review, payment tracking, search and filters, archive and restore, CSV export, and dashboard summaries.

The public demo explains and previews the workflow with sample invoices. Persistent workspaces and AWS-backed extraction depend on the configured backend and are not independently verified here.

Technical decision

Keep original documents beside editable records and require human confirmation before an invoice affects totals; AWS extraction remains dependent on backend configuration.

Open demo

My role

Product direction
UX/UI design
Full-stack development
AWS integration

Technology

NuxtVueTypeScriptTailwind CSSNode.jsAWS

AI-assisted job search workspace / 2026

Zentail

Launched · Early release

One dashboard to organize a job search: track applications and stages, analyze job fit with AI, compare resume results, and see overall progress.

Current scope: Built with my co-founder over four weeks, inspired by our own placement-season job searches. Available to try and actively being improved.

Next.jsReact

Open project

AI customer support agent frontend / 2026

Relay

In progress · Frontend prototype

A responsive support workspace for order tracking, refunds, return policies, product search, and simulated tickets. Conversation context and an activity panel expose tool inputs and results.

Current scope: Working frontend with simulated responses and tool calls.
Planned: LLM integration and a FastAPI backend.

ReactTypeScriptTailwind CSSshadcn/ui

AWS✳BEDROCK✳LLMS✳PYTHON✳TYPESCRIPT✳NUXT✳VUE✳NODE✳RAG✳AWS✳BEDROCK✳LLMS✳PYTHON✳TYPESCRIPT✳NUXT✳VUE✳NODE✳RAG✳

( 03 / AWS capability → evidence )

I’m building toward customer-facing cloud roles by connecting each AWS service to a concrete application decision—not a logo list.

Evidence below reflects the described Contexta implementation; public-demo availability is stated separately.

Identity & access

Amazon Cognito · IAM

Contexta’s described implementation uses authenticated sessions, protected APIs, and permission-aware retrieval.

Serverless APIs

API Gateway · AWS Lambda

Contexta separates the Nuxt client from the retrieval workflow through a protected API and Python Lambda functions.

Storage & application data

Amazon S3 · DynamoDB

The architecture separates private document storage from conversation history, feedback, and application state.

Generative AI & RAG

Amazon Bedrock · Bedrock Knowledge Bases

Contexta retrieves relevant company material before generating an answer and returns citations for verification.

Operations & evaluation

CloudWatch · Guardrails · RAG evaluation

The project scope includes latency monitoring, guardrail support, user feedback, and a versioned evaluation suite.

( 04 / Contact )

Let’s discuss
software.

Let’s discuss software, cloud, and customer problems. I’m seeking opportunities in software development and cloud solutions, with a particular interest in customer-facing work and generative AI.

Email me GitHub

I turn businessneedsintopractical software.

AWS Certified AI Practitioner

Built tomatter.

Contexta

Floorsir

Invoxa

Zentail

Relay

Identity & access

Serverless APIs

Storage & application data

Generative AI & RAG

Operations & evaluation

Let’s discusssoftware.

Built to
matter.

Let’s discuss
software.