COMPUTER ENGINEERING · ARTIFICIAL INTELLIGENCE · CYBERSECURITY · FLUTTER
From idea to working prototype; from research to clear user experience.
“Teknolojiyi göstermek için değil, doğru sonucu üretmek için kullanıyorum.”WHO I HELP
I assess scope first for websites, mobile UI, technical prototypes, software testing, student mentoring and research support.
Portfolio, CV site, project mentoring, code review and technical presentation support.
Testable prototypes, landing pages, forms, basic automation and launch support.
Technical research, test scenarios, documentation and digital workflow support.
VERIFIABLE PROFILE
My academic base, primary career direction and applied product experience are presented as distinct, honest tracks.
Algorithms, data structures, OOP, databases and software engineering.
Human review, data boundaries, secure software and defensive thinking.
Modular UI, responsive implementation, testing, debugging and performance.
Problem and risk first; measurable output and traceable delivery next.
01 · PROFESSIONAL PROFILE
Fourth-year Computer Engineering student at TED University and early-career developer. The main direction is AI systems, machine-learning foundations, cybersecurity and secure software; Flutter, APIs, web, testing and debugging form a separate applied product-development track.
I turn academic foundations in algorithms, data structures, software engineering, databases and programming languages into practical product decisions.
During my 2025 internship in the software department of an established health-technology organisation in Türkiye, I worked on modular Flutter/Dart UI, REST APIs, responsive implementation from Figma, scenario testing, debugging and performance improvement.
I develop AI prototypes and TryHackMe-based security practice through principles of data responsibility, human approval, testability and clear usage boundaries.
02 · FOCUS & CAPABILITIES
Areas I have applied, studied and continue to develop actively.
LLM workflows with explicit data boundaries, human approval points, failure cases and measurable small prototypes.
Modular Flutter/Dart components, REST API states and consistent interaction across screen sizes.
Authorised TryHackMe labs complemented by input validation, least privilege, data-flow review and safe failure behaviour.
Reproducing issues, isolating causes and documenting user flows, edge cases and prioritised improvements.
Connecting REST contracts, validation, loading/error states and accessible client-server interfaces.
Assessing trustworthy sources with assumptions, comparisons and risks for technical and non-technical readers.
Flutter, Dart, API integration, responsive UI
AI applications, Python, cloud and system design
TryHackMe, web security and defensive thinking
Java, C, SQL, algorithms, data structures, OOP
ACTIVE LEARNING
Not claims of expertise—areas I am actively studying to understand their practical boundaries.
I am studying where agentic workflows create value and how their outputs can be evaluated.
LEARNING · APPLYINGI am bringing input validation, access boundaries, safe failure and defensive thinking earlier into development.
LEARNING · APPLYINGI am reinforcing networking, Linux, web-security and defensive-thinking foundations through hands-on labs.
LEARNING · APPLYINGI am steadily developing Python, system design, evaluation and safe-AI boundaries.
LEARNING · APPLYING03 · SELECTED WORK
Each card explains what I did and where the work currently stands.
LIVE WEB PRODUCT
A technical profile, services, CVs, pricing and thirteen locale routes needed one trustworthy product experience.
I led information architecture, interface design, React/TypeScript implementation, multilingual content, accessibility and delivery.
I used transparent scope, a local rule-based assistant, indexable locale routes and provider-independent email options.
A working responsive, theme-aware and indexable product. The image is captured from the real interface.
Live demo + real interface capture + browser QA
CS50x CAPSTONE
A web product was needed to explain password strength for different platforms and retain analysis history.
I worked on Flask routing, SQLite history, responsive HTML/CSS/JavaScript UI and persistent theme preference.
To respect CS50 academic-honesty rules, the public repository documents scope and features while the solution source remains private.
A full-stack capstone whose features and technology are publicly verifiable in the GitHub portfolio.
Public GitHub documentation; solution source kept private
FLUTTER UI IMPLEMENTATION
The visible, validation and error states of a login screen needed one clear mobile interaction flow.
I implemented responsive Flutter/Dart layout, controller lifecycle, StatefulWidget state, password visibility and demo error feedback.
This is a UI prototype. Apple/Google buttons are visual representations; there is no OAuth, backend, real account or production authentication workflow.
A bounded mobile UI implementation evidenced by source, emulator output and three visible states.
Original user-supplied Flutter work image · IDE/emulator evidence · documented source scope
I’m Seyid Berat Ergün, a fourth-year Computer Engineering student at TED University and an early-career developer. My primary direction is AI systems, machine-learning foundations, cybersecurity and secure software. Separately, I build applied product experience in Flutter/Dart, REST APIs, responsive web interfaces, testing and debugging.
ENGINEERING LABS / GITHUB PORTFOLIO
Algorithms
Graph algorithm laboratory implementing Prim, Kruskal and Union-Find with deterministic tests and complexity documentation.
Cybersecurity
A bounded local rule engine that routes text to ALLOW, REVIEW or BLOCK with stable reason codes.
AI / ML
Dependency-free toolkit for evaluating binary-classifier outputs. Calculates confusion-matrix metrics, Brier score, threshold sweeps and calibration bins so model behavior can be inspected without hiding the calculations behind a training framework.
Compact RAG evaluation workbench for recorded retrieval experiments. Calculates Precision@K, Recall@K, MRR, nDCG and a lexical-groundedness proxy without requiring a live model or vector database.
Local rule-based guardrail prototype for reviewing recorded text before an AI workflow. Flags prompt-injection, secret-like assignments, email exposure and exfiltration-style requests with explainable ALLOW, REVIEW or BLOCK decisions.
Lightweight model-input drift monitor for comparing reference and current datasets. Measures numeric PSI, categorical distribution change and missing-rate differences with deterministic local reports.
Local SOC-style correlation engine for recorded authentication and MFA events. Groups events across configurable time windows to surface failure bursts, success-after-burst patterns and repeated MFA denials with pseudonymized actor IDs.
Defensive API policy lab for reviewing recorded request and response metadata. Checks method and body limits, CORS signals, request paths, content types and local rate-limit behavior without sending traffic to a live target.
Small contract-testing utility for recorded JSON API responses. Validates expected status codes, required fields, primitive types and unexpected fields with deterministic error paths for repeatable offline regression tests.
Local configuration scanner for catching common privacy and security mistakes before code is shared or deployed. Detects hardcoded secrets, debug mode, permissive CORS and non-local plaintext HTTP references without printing matched secret values.
Threshold-analysis toolkit for recorded binary-classifier scores. Shows how confusion-matrix counts, precision, recall, F1 and accuracy change as the decision threshold moves, with a lightweight self-contained HTML report.
Interactive browser visualizer for BFS and Dijkstra pathfinding. Build walls and weighted cells, compare shortest-path behavior on the same grid and test the core algorithms independently from the interface.
Local security-event dashboard for authentication-failure analysis, time-windowed event grouping and privacy-conscious monitoring of recorded security events.
Local log-triage utility for access-log-style data. Redacts common credential-like values and email addresses, summarizes log patterns and highlights authentication-failure bursts without sending logs to an external service.
Explainable Turkish and English phishing-risk analyzer for messages and URLs. Combines social-engineering language signals with lexical URL checks and returns LOW, MEDIUM or HIGH guidance without visiting links or uploading submitted content.
04 · EDUCATION & EXPERIENCE
EDUCATION
Fourth-year student; expected graduation June 2027. Algorithms, software engineering, data structures, databases, programming languages and digital logic.
INTERNSHIP
Applied work in its software department across Flutter/Dart, REST APIs, responsive UI from Figma, scenario testing, debugging, performance and refactoring.
TECHNICAL TEAM
Live broadcast and recording infrastructure, audio equipment, incident response and process standardisation.
SELF-DEVELOPMENT
Guided labs in networking, Linux, web security and security mindset.
LEADERSHIP
Ankara rankings, team coordination, mentoring and decision-making under pressure.
CERTIFICATES & ACADEMIC TRAINING
Implementation discipline, test strategy, defect thinking and software-quality foundations.
Requirements, architecture choices, project planning, risk and delivery communication.
System boundaries, use cases, class/sequence models and clearer design reasoning.
Computer-science fundamentals through algorithms, memory, data structures and practical problem solving.
Deliberate practice in Java collections, complexity, reusable patterns and coding questions.
Cloud service concepts, developer workflow, deployment thinking and platform fundamentals.
Purposeful programming, algorithms and problem decomposition with Java-based exercises.
Object-oriented modelling, encapsulation, inheritance, interfaces and maintainable Java design.
Responsible use of AI in academic production, institutional processes and human-reviewed output.
Names and dates match the original résumé. LinkedIn is used for personal credential links; no credential ID is invented on this site.
05 · SERVICES & PRICING
Early-career starting prices that stay accessible without undervaluing preparation and delivery. TRY, approximate USD and EUR are shown together.
Clarify technology, risk, budget and the first steps for your idea, coursework plan or digital need.
One-to-one explanation, guidance, code review, testing and debugging support for students who want to understand their work.
Support to make an existing project clearer, more stable and ready to deliver.
A controlled, measurable and human-approved first AI solution for individuals or small teams.
A fast, responsive and professional one-page experience for a brand, service or project.
Responsive, fast and maintainable multi-page websites for personal or organisational needs, including interface work, basic forms/integrations and launch preparation.
Make a mobile idea visible and testable before stakeholder or user conversations.
Request a scope review for a platform, integration, prototype or unlisted need. Work outside the early-career reference ranges receives a tailored budget and delivery plan.
BUDGET GUIDE
These are indicative scope ranges, not fixed prices or offers.
| Service level | TRY | USD / EUR | Typical scope |
|---|---|---|---|
| 01Technical consultation and practical roadmap | TRY 5,000 – TRY 10,000 | USD — / EUR — | Clarify technology, risk, budget and the first steps for your idea, coursework plan or digital need. |
| 02Education & Project Mentoring | TRY 5,000 – TRY 15,000 | USD — / EUR — | One-to-one explanation, guidance, code review, testing and debugging support for students who want to understand their work. |
| 03Code Review, Testing & Debugging | TRY 5,000 – TRY 20,000 | USD — / EUR — | Support to make an existing project clearer, more stable and ready to deliver. |
| 04AI Workflow & Prototype | TRY 15,000 – TRY 75,000 | USD — / EUR — | A controlled, measurable and human-approved first AI solution for individuals or small teams. |
| 05Portfolio & Landing Page | TRY 10,000 – TRY 45,000 | USD — / EUR — | A fast, responsive and professional one-page experience for a brand, service or project. |
| 06Website Development | TRY 25,000 – TRY 150,000 | USD — / EUR — | Responsive, fast and maintainable multi-page websites for personal or organisational needs, including interface work, basic forms/integrations and launch preparation. |
| 07Interactive Flutter Prototype | TRY 20,000 – TRY 90,000 | USD — / EUR — | Make a mobile idea visible and testable before stakeholder or user conversations. |
| 08Need something different? | TRY 30,000 – TRY 560,000+ | USD — / EUR — | Request a scope review for a platform, integration, prototype or unlisted need. Work outside the early-career reference ranges receives a tailored budget and delivery plan. |
The final fee is confirmed in writing after screens, integrations, content, delivery time and revisions are known.
USD and EUR are approximate conversions from base TRY prices using the dated ECB reference; they are omitted when unavailable. The accepted quote is fixed in the selected currency. Domains, hosting and paid services are separate. Mentoring focuses on explanation, planning, review and debugging; the student remains academically responsible for submitted work.
Displayed amounts are estimated price ranges. Final scope, delivery plan and fee are determined after a written proposal.
WORKING & QUOTE TERMS
Website information is not a binding order or contract. Work begins only after scope, total price, tax/document treatment, delivery, revisions and responsibilities are accepted in writing.
A form or email is only an enquiry. The service becomes definite after the written quote or contract is accepted.
Deliverables, timing, communication, acceptance criteria and responsibilities are written in the quote.
Taxes and third-party costs such as domains, hosting, store accounts, licences and paid APIs are shown before acceptance; the legally required document is issued.
Included revisions are defined; out-of-scope requests require separate timing and fees.
The client confirms the right to provide supplied text, media, data, accounts and access.
Unauthorised access, harmful use and producing assessed work on a student's behalf are outside scope; mentoring supports learning.
For consumer work, pre-contract information, withdrawal rights and any request to start before the withdrawal period ends are handled for that project.
Source files, licence, payment schedule, cancellation/refund terms and support are defined per proposal.
This is general information, not legal advice or a standalone contract. Professional legal review is recommended before commercial use.
06 · PROCESS
You know the boundaries, deliverables and next step at every stage.
Clarify the user, problem, budget and delivery expectation.
Document deliverables, timeline, fee and revision boundaries.
Review a visible draft before development within the approved scope.
Test core scenarios and deliver files, notes and next steps.
07 · CURRENT NOTES
A curated selection from primary sources such as OWASP, FIDO Alliance and Flutter. Editorial summaries guide the reader; the complete article always opens on its canonical source.
Short original notes make my learning, implementation boundaries and engineering decisions visible.
Beyond impressive demos: a concise guide to automation, tool use, human approval and security boundaries in real scenarios.
Read full noteA plain-language look at passwordless sign-in, what it improves for users and which common security problems it reduces.
Read full noteWhy architecture, state, APIs, testing and performance become critical when moving beyond demo screens.
Read full noteA good testing approach does more than check whether an application crashes. It helps reveal the paths users can take, the effect of unexpected input and whether a change breaks previously working behaviour. Even in small projects, scenario-based thinking makes problems visible before they become expensive to solve.
A model producing an answer is not enough to prove success. Consistency, hallucination risk, instruction following, sensitive-information boundaries and contribution to the real task should be evaluated together. In AI products, evaluation is one of the engineering steps that separates an impressive demo from genuinely reliable use.
A local-first approach uses narrow, deterministic and verifiable paths before sending an unanswered question to a model. This technical note examines routing, external-source trust boundaries, server-side fallback and a data flow designed without persistent chat history through the current SBE Assistant architecture.
Read full noteOWASP GenAI Security ProjectAuthor / contribution: 100+ industry contributors
An open industry framework for critical risks in AI systems that plan, use tools and act—covering goal hijack, identity and permission boundaries, human oversight and safe failure.
UK National Cyber Security CentreAuthor / contribution: Dave Chismon
NCSC's assessment uses real-world attack techniques to compare passkeys and FIDO2 credentials with passwords and traditional MFA across phishing, credential reuse and account recovery.
Flutter DocumentationAuthor / contribution: Flutter team
Flutter's living release page brings together Flutter 3.44, updated learning paths, accessibility guidance, tooling, breaking changes and official migration resources.
08 · FREQUENTLY ASKED QUESTIONS
Scope, communication, academic guidance, security approach and pricing.
It is an automated information assistant that instantly matches typed questions to services, pricing, AI, cybersecurity, experience and contact information. It stores no messages, does not pose as a human, and directs detailed requests to Gmail.
Every engagement uses a clear scope, written deliverables, regular communication, an early visible draft and basic testing. Early-career pricing keeps the process agile while this structure protects professional quality.
We first break down the question and learning goal. I then support the student through concept explanation, algorithm planning, example approaches, code review, debugging and presentation preparation.
My TryHackMe labs and secure-software studies support basic risk checks, test scenarios, input validation, failure-flow review and security-awareness recommendations. I can also help identify the right specialist service when formal assessment is required.
We can communicate directly in Turkish and English. Start through Gmail and review my work on GitHub and LinkedIn. The interface is also available in Spanish, Italian and German.
Domain selection, DNS, hosting setup, basic SEO, performance checks and short post-launch support can be added according to scope.
TRY is the base price. USD and EUR use the dated ECB reference for approximate conversion, not tick-by-tick rates; unavailable references are omitted. Once screens, integrations, content, timing and revisions are clear, a written quote is fixed in the selected currency.
09 · CONTACT
Share your idea, problem or goal. I will be honest about fit and help define a practical first step.
You can contact me in Turkish or English.
Detailed education, experience and capabilities.
TR · Download Turkish CV ↓EN · Download English résumé ↓