Ali Cakil Book a call

Portrait of Ali CakilAli Cakil · Software engineer in Madrid

I build AI agents and LLM features into real products.

Twelve years of .NET, with AWS and Angular alongside. I founded Vitena, a SaaS for health professionals that runs in production with its own LLM layer, tool-calling agents and a public MCP server.

Book a call See my work CV (PDF)
  • AWS Certified Developer
  • Remote work with international teams
  • English C1 · Turkish native
Vitena's dashboard for a practitioner (demo account)
Vitena's client portal on a phone (demo account)

AI assistantbuilds meal plans in chat

Public MCP serverin the official registry

Teams and clients I have worked with

  • Jabil
  • ESW
  • Fine Hygienic Holding
  • Stow Group
  • NTDCO
  • Mebant
  • Kasih Food
  • Euro Gıda
  • Genesis

Selected work

Products I have built end to end, from the database to the AI layer.

A live product I founded, a freelance app, and systems that have run warehouses and enterprise teams for years.

Live productFounder · Nov 2025 to present

Vitena

A practice-management platform for health and wellness professionals, with a public directory where clients find and book them. It runs in production with real users, in 7 languages, on the web and as an Android app.

  • 7languages, on the web and Android
  • 20+LLM task types in production
  • 3AI agents that talk to users
  • 84–98%of prompt tokens served from cache
  • .NET 10 on AWS Lambda
  • Angular with SSR
  • PostgreSQL, schema per tenant
  • AWS CDK
  • Android app

vitena.care MCP Registry For AI assistants Android app

Vitena's meal-plan builder with the AI assistant panel open (demo account)
The AI assistant drafts a meal plan from the practitioner's own recipes and foods; the nutrition values are computed in code, not by the model.
A practitioner's public page on Vitena with a booking button (demo account)
Every practitioner gets a public page with online booking, listed in Vitena's directory.

Built end to end, by one engineer

Product, backend, front end, infrastructure and the AI side: from the first commit to real users in production.

  • .NET 10 API on AWS Lambda with multi-tenant PostgreSQL
  • Angular with server-side rendering, 7 languages, an Android app
  • Infrastructure as code with AWS CDK and CodePipeline
  • An LLM layer with agents, evals and cost metering

The AI side

All of it is live and in daily use, not a demo.

  • Provider-agnostic LLM layer

    One completion interface in .NET with an adapter per provider; model and cost policy per task type, every call metered with its real cost.

  • Agents in production

    A tool-calling agent loop behind in-app support, a public pre-sales assistant and an admin agent on read-only SQL with hard limits.

  • Retrieval without embeddings

    A document map in the system prompt plus a read_doc tool, with the prompt ordered so most input tokens come from cache.

  • 20+ LLM task types

    Meal plans, session-note drafts, document reading with vision, voice-to-record capture, and triage of email and feedback.

  • Tracing and evals

    Tracing at the adapter layer, a spend report by task type, and a black-box eval harness that runs whenever a prompt changes.

  • Public MCP server

    A read-only MCP server over Streamable HTTP, listed in the official MCP Registry as care.vitena/directory.

  • Readable to AI assistants

    Server-rendered pages, JSON-LD, a sitemap and llms.txt, with server-side analytics for AI crawlers.

  • Coding-agent pipeline

    Claude Code sessions work through a task step by step: one plan approval up front, a fixed stop list, deploys behind a permission gate.

More projects

Freelance and client work, before Vitena.

Illustration: an event discovery app and a QR-code ticket on two phones Illustration

Freelance · 2022 – 2024

Social events and ticketing app

A listing app where people discover events and buy tickets, built as a serverless application on AWS.

  • Ticket sales with Stripe and QR-code tickets for check-in
  • LLM-based parsing of event content and scraping from other platforms
  • Location-based search with Google Maps, Google and Facebook sign-in
  • Lambda
  • Step Functions
  • Cognito
  • PostgreSQL
  • .NET
  • Node.js
  • Angular
Illustration: an automated high-bay warehouse with a warehouse management dashboard on a tablet Illustration

Genesis AS · 2015 – 2022

Warehouse management systems

Custom WMS and automated-storage software for logistics and manufacturing, from pre-sales to go-live and on-site maintenance.

  • Barcode-based inventory tracking and quality control
  • Algorithms for route planning and truck loading
  • Windows and web applications on ASP.NET

For clients including Fine Hygienic Holding (Abu Dhabi plant), Stow Group, NTDCO, Mebant, Kasih Food and Euro Gıda.

  • C#
  • ASP.NET
  • SQL Server
  • WMS
  • AS/RS
Illustration: an event-driven serverless architecture with an API gateway, functions, queues and an event bus Illustration

Jabil · 2022 – 2024 · remote

Enterprise web apps on AWS

Senior .NET and AWS developer in a Scrum team for Jabil, a global manufacturer, including the backend of a meeting and action-management app.

  • Event-driven services on Lambda, Step Functions, EventBridge, SQS and SNS
  • Fargate and DynamoDB; ETL with AWS Glue and Athena
  • Deployments automated with AWS SAM, CloudFormation and Azure Pipelines
  • C#
  • .NET
  • AWS
  • DynamoDB
  • AWS SAM

What I do

Four ways I can help your product

Each one is something I have built and run on a live product.

AI agents and tool calling

Agent loops that work on your own data and APIs through tools, with guardrails: read-only tools, limits on what a query can touch, and approval before any irreversible step. Also MCP servers, so outside assistants can use your product.

LLM integration in .NET

An LLM layer inside an existing .NET backend: one interface with an adapter per provider, model and cost policy per task type, usage metering, tracing, and evals that run whenever a prompt changes.

.NET backends on AWS

Serverless and event-driven services in C# and ASP.NET Core on Lambda, API Gateway, SQS and Step Functions, with infrastructure as code and CI/CD. Angular on the front end when the job needs full stack.

Agentic development with Claude Code

Coding-agent workflows on a production codebase: work planned up front and approved once, a fixed list of cases where the agent must stop and ask, and production deploys behind a permission gate.

Experience

Twelve years of software that runs real businesses

  1. Nov 2025 to present

    Founder & Software Engineer · Vitena

    Building the platform above: backend, front end, infrastructure and the AI side.

  2. Nov 2024 to Nov 2025

    .NET Backend Developer · eCommerce Platform · Madrid and Málaga, hybrid

    Microservices-based e-commerce on Azure Cosmos DB with DDD, MediatR and CQRS. Code quality with xUnit, in-memory integration tests, Sonar and mutation testing with Stryker.NET.

  3. Jun 2022 to Oct 2024

    Senior .NET & AWS Developer · Jabil Inc. · USA, remote

    Web applications in a Scrum team on AWS: API Gateway, EventBridge, Step Functions, Lambda, Fargate, SQS, SNS and DynamoDB. Deployments automated with AWS SAM, CloudFormation and Azure Pipelines; ETL with AWS Glue and Athena.

  4. Jan 2022 to Nov 2024

    Freelancer · social events web application · remote

    The events and ticketing app above.

  5. Apr 2015 to May 2022

    Full Stack Developer · Genesis AS · Türkiye, on‑site

    Custom warehouse management systems for logistics: barcode-based inventory tracking, quality control, and algorithms for route planning and truck loading. Windows and web ASP.NET applications, from pre-sales to on-site maintenance.

Skills

AI
LLM integrationAI agentsTool callingMCP serversPrompt engineeringRAGEvals and tracingOpenAI APIClaude Code
Backend
C#.NET 8 and 10ASP.NET CoreMinimal APIsEF CoreRESTCQRSDDDMicroservicesEvent-drivenxUnitTDD
Cloud and delivery
AWS LambdaAPI GatewaySQSStep FunctionsAWS SAMTerraformAWS CDKCI/CDGitHub ActionsAzure DevOpsDockerKafka
Data
PostgreSQLSQL ServerDynamoDBCosmos DBETL
Frontend
AngularTypeScriptJavaScript

Credentials

Contact

Have a project in mind? Let's talk.

Tell me what you are building and where an agent or an LLM feature should fit. A 30-minute intro call on Google Meet is the quickest start; email works just as well.

Book a 30-min call alicakil78@gmail.com

LinkedIn WhatsApp Telegram CV (PDF)

Based in Madrid, Spain · Central European Time