Portfolio

AsliddinErgashev

Full Stack AI Engineer | AI Integration Specialist

Bringing AI innovation into real-world business solutions. Specializing in LLM-powered document intelligence, RAG systems, and production-ready AI applications.

View Projects→Get in Touch

Otto-Friedrich-Universität Bamberg

Asliddin Ergashev
SCROLL

About Me

4+

Years in AI/ML

15+

Production Projects

3

Major Companies

Full Stack AI Engineer
& Integration Specialist

I'm Asliddin — a Full Stack AI Engineer specializing in bringing LLM-powered solutions into production. Currently at Trench Group, I develop AI systems that transform how technical sales teams work with complex documents.

My work on Project NOVA's Spark — an LLM-powered document intelligence platform — extracts, structures, and compares customer specifications using GPT-4/5, RAG systems, and advanced prompt engineering. I've processed 100-500 page documents across heterogeneous formats (PDF, Word, PowerPoint, Excel, images).

Previously at Siemens Healthineers and Brose Group, I automated technical workflows, built DevOps pipelines, and developed production systems with FastAPI, React, and PostgreSQL.

Tech Stack

LangChainAzure OpenAIGPT-4/GPT-5RAG SystemsFastAPIReact/Next.jsTypeScriptDocker

Research Interests

  • →LLM-Powered Document Intelligence
  • →Retrieval Augmented Generation (RAG)
  • →AI Agents & Agentic Systems
  • →Prompt Engineering & Evaluation
  • →Human-AI Interaction Research

Selected Projects

01

AI-Based Adaptive Learning Assistant

Bachelor thesis research prototype demonstrating three design principles for intelligent tutoring systems: personalized adaptive content based on student materials, context-aware conversational interface grounded in documents, and transparent AI reasoning with visible sources. Features RAG pipeline with semantic search, document chunking, quiz tracking, and AI-powered adaptive question generation.

Next.jsTypeScriptSupabasepgvectorGoogle Gemini APIRAG
View project →
02

Full AI Engineer Portfolio

Modern, responsive portfolio website showcasing projects, research, and professional work. Built with Next.js 16 and TypeScript, featuring smooth animations, dark theme, optimized font loading, and deployed on Vercel with custom domain integration.

Next.js 16TypeScriptTailwind CSSFramer MotionVercel
View project →
03

MeinHelfer

Full-stack managed marketplace for booking trusted service providers (moving, cleaning, gardening). Features async FastAPI backend with JWT authentication, PostgreSQL database, and Dockerized production deployment.

Next.js 15TypeScriptTailwind CSSFastAPIPostgreSQLDocker
View project →
04

Gut Guardian

Gamified mobile app helping users identify and manage food intolerances through structured challenges, progress tracking, and habit loops over a 14-day programme. Built for Viatolea Challenge.

React NativeExpoTypeScriptC#
View project →
05

Local LLM Chat App

Full-stack app enabling users to interact with locally running LLMs (Mistral/Phi via Ollama) through a web interface — no API keys or cloud required. FastAPI backend handles prompt forwarding; vanilla JS frontend renders responses in real time.

FastAPIPythonOllamaMistralHTML/CSS/JS
View project →
06

Uzbek-German AI Teacher

AI-powered language learning assistant for Uzbek-German translation and language practice, leveraging LLM technology for interactive learning experiences.

PythonLLMAI
View project →
07

Jobwiese.com

Bilingual (DE/EN) marketing website for a German job portal with locale-based routing and server-side transactional email delivery via Resend API.

Next.js 16TypeScriptTailwind CSSResend APINetlify
View project →
08

VEN-Marketing.com

Zero-dependency single-page marketing site with animated canvas, live counters, scrolling ticker, and bilingual (DE/EN) toggle. Built with vanilla technologies for maximum performance.

HTML5Vanilla CSSJavaScriptPHP
View project →
09

California Housing Price Prediction

Machine learning project modeling housing costs for a California construction company using regression techniques and feature engineering.

PythonJupyter Notebookscikit-learnPandasNumPy
View project →
10

Data Science & AI Portfolio

Collection of data science and AI projects completed in Google Colab and Jupyter Notebook, covering various ML/AI techniques and algorithms.

PythonJupyter NotebookMLData Science
View project →

Academic Research

Research-in-Progress Paper·2024·Research-in-Progress

Understanding Human–AI Relationships: A Conceptual Framework for Designing Trustworthy AI Companions

Supervised by Prof. Dr. Milad Mirbabaie · Chair of AI Engineering in Companies · University of Bamberg

Bachelor Thesis·2025·In progress

Designing an AI-Based Adaptive Learning Assistant: A Design Science Research Approach

Otto-Friedrich-Universität Bamberg · Chair of AI Engineering in Companies

Research within the Design Science Research paradigm — human–AI interaction and AI-augmented learning systems.

Writing

Short write-ups on things I've learned.

May 2025·AI

What Design Science Research Actually Means for AI Projects

DSR isn't just a methodology label — it changes what you build, why you build it, and how you know when you're done.

Coming soon
Mar 2025·Unity

Gamification Beyond Points: What I Learned Building Viatolea

Most gamification advice stops at badges and streaks. Designing for a sensitive health context forced me to think about motivation differently.

Coming soon
Jan 2025·Web Dev

The Hidden Cost of a Fast Dev Server

Turbopack and Vite make iteration feel instant, but they also make it easy to forget what your app actually costs at runtime.

Coming soon

Video

Videos from events, conferences, and moments worth sharing.

2025·Event

First event video

Add a youtubeId to this entry to activate the thumbnail and link.

Coming soon
2025·Event

Second event video

Conferences, demos, presentations — anything worth sharing from the events you attend.

Coming soon

Get In Touch

Let's buildsomething great.

Open to freelance work, research collaborations, and interesting conversations. Fill out the form or reach out directly.

Request a Service

© 2026 Asliddin Ergashev

asliddin-ergashev.com