Full-Stack AI Engineer · Web & Agentic I build products that work beyond the demo.
Available for new work.
Full-stack engineering for software that needs to keep working after launch.
- Multi-agent systems
- MCP + integrations
- RAG pipelines
- React & Next.js
Ahmed builds production AI systems and the full-stack products around them — LangGraph, MCP and RAG on the back end, React and Next.js on the front.
More than algorithms.
Hi, I’m Ahmed Islam, a Full-Stack AI Engineer with a passion for building efficient, scalable, and intelligent systems. With 2+ years of experience in the field, I work across the whole build — the model, the API behind it, and the interface people actually use.
I believe that AI is about more than just complex algorithms, it’s about solving tangible problems and creating intuitive, data-driven experiences for users.
Whether I’m working on an NLP model, a data extraction pipeline, a generative AI application or a React front end, I bring my commitment to engineering excellence and user-centered solutions to every project. I look forward to the opportunity to bring my skills and passion to your next challenge.
Where I’ve been.
Live SaaS platform
Padelos
Associate AI Engineer
Jan 2026 — Present
- Build LLM-powered agents and multi-agent workflows on the Padelos SaaS platform with LangChain and LangGraph, automating multi-step production processes.
- Developed a Model Context Protocol (MCP) server that exposes platform data and business logic to AI agents through structured tool-calling interfaces.
- Engineer and version system prompts for PadelGPT, the platform assistant, and built a quality-evaluation pipeline that grades responses against a rubric and generates structured Excel and CSV reports.
- Manage agent memory with vector stores (FAISS, ChromaDB) for retrieval, and Redis and PostgreSQL for short-term and persistent state.
- Built n8n pipelines that extract, transform, and report conversation data from Azure Cosmos DB across player and admin schemas.
- Migrated the platform chatbot from Azure to AWS EC2 and shipped services with Docker and Bitbucket Pipelines CI/CD, adding CloudWatch-based observability.
- LangGraph
- LangChain
- MCP
- FAISS
- ChromaDB
- Redis
- PostgreSQL
- n8n
- AWS EC2
- Docker
The full stack.
Agents and LLM systems at the centre, with the ETL pipelines, databases and cloud plumbing that keep them running in production.
70 / 70 tools
- LangChain
- LangGraph
- MCP
- RAG
- Tool calling
- Multi-agent orchestration
- Prompt engineering
- Generative AI apps
- OpenAI
- Anthropic Claude
- Google Gemini
- LLaMA
- Mistral
- Ollama
- Hugging Face
- Groq API
- Fine-tuning
- scikit-learn
- TensorFlow
- PyTorch
- Deep Learning
- CNNs
- U-Net
- YOLO
- NLP
- Computer Vision
- Transfer Learning
- FastAPI
- Flask
- FastMCP
- ETL / ELT pipelines
- n8n
- REST APIs
- Socket.io
- Postman
- pandas
- React
- Next.js
- HTML
- CSS
- Tailwind CSS
- Bootstrap
- AWS EC2
- CloudWatch
- Azure App Service
- Google Cloud
- Docker
- Kubernetes
- Bitbucket Pipelines
- Git
- GitHub
- Vercel
- Netlify
- OpenTelemetry
- Langfuse
- FAISS
- ChromaDB
- Redis
- PostgreSQL
- MySQL
- SQLite
- MongoDB
- Firebase
- Python
- SQL
- JavaScript
- C++
- Power BI
- Streamlit
- Gradio
What I deliver.
Seven ways I get hired, from a single agent to the whole product around it. Scope one or chain them — most projects end up being two or three.
Multi-agent workflows that do real work instead of demoing well.
- Agent graph design — state, routing, retries, human-in-the-loop
- Tool surface over your existing APIs and business logic
- An MCP server so any client can call those tools
- Rubric-graded evaluation before anything reaches users
- LangGraph
- LangChain
- MCP
- Python
Answers grounded in your own documents, with the retrieval measured.
- Ingestion and chunking pipeline for your corpus
- Vector store selection, indexing and hybrid retrieval
- Multimodal retrieval across text and images
- Retrieval scoring, so you can see what it actually found
- RAG
- FAISS
- ChromaDB
- Hugging Face
Fast, responsive builds in React and Next.js — no AI required.
- Marketing sites and landing pages that load fast and rank
- React and Next.js applications with routing, auth and real state
- Dashboards and admin panels over your existing API
- Responsive from 390px up, accessible, and yours to hand to any developer
- React
- Next.js
- Tailwind CSS
- Vercel
The whole product around the model — API, database, and the front end people use.
- FastAPI or Flask backend with authentication and rate limiting
- Schema design on PostgreSQL, MongoDB or Redis
- React or Next.js front ends with streaming chat and live dashboards
- Internal tools and quick prototypes with Streamlit or Gradio
- FastAPI
- PostgreSQL
- Tailwind CSS
- Streamlit
The unglamorous plumbing that moves your data where it needs to be.
- n8n workflows connecting the tools you already pay for
- ETL / ELT pipelines with scheduling and retries
- Large-scale scraping and data ingestion
- Scheduled Excel and CSV reporting off live data
- n8n
- ETL / ELT pipelines
- pandas
- REST APIs
Custom models for when a general-purpose API is the wrong tool.
- Image classification, detection and segmentation
- Training on your data, with augmentation for class imbalance
- NLP classification and extraction
- Honest evaluation — accuracy, Dice, confusion matrices
- PyTorch
- TensorFlow
- U-Net
- Computer Vision
Getting it live, and knowing the moment it stops working.
- Containerisation and CI/CD pipelines
- Deployment to AWS or Azure
- Tracing and token/cost tracking on every LLM call
- Alerting and dashboards on the metrics that matter
- Docker
- AWS EC2
- Langfuse
- OpenTelemetry
In their words.
Five recommendations from LinkedIn, opened at their sharpest line. Pick any one to read it in full, exactly as it was written.
Working with Ahmed felt less like hiring an engineer and more like gaining a partner who genuinely cared about the outcome.
1 With a strong grasp of complex concepts, he not only shares insightful opinions but also motivates others.
2 He has developed solid expertise in Artificial Intelligence and Machine Learning, not just in theory but also through practical implementation.
3 His ability to troubleshoot issues and find efficient workarounds played a crucial role in the success of our work.
4 He was amazing to work with and really compromising personality ❤
5
Who said it
Working with Ahmed felt less like hiring an engineer and more like gaining a partner who genuinely cared about the outcome. He took a rough idea I had around AI and turned it into a polished, production-ready solution — explaining every decision in plain language along the way. What stood out most was his reliability: clear updates, zero surprises, and models that actually performed in the real world. If you want someone who blends deep AI/ML expertise with the rare ability to just get things done, Ahmed is the one. I’d work with him again in a heartbeat.
It has been a pleasure working with Ahmed. His collaborative nature makes every project enjoyable and productive. With a strong grasp of complex concepts, he not only shares insightful opinions but also motivates others. I have had the opportunity to work with him on AI and deep learning tasks, and his expertise has consistently contributed to our success.
I had the pleasure of studying alongside Ahmed at the University of Haripur. During that time, I witnessed firsthand his strong dedication to learning and his passion for technology. He has developed solid expertise in Artificial Intelligence and Machine Learning, not just in theory but also through practical implementation. Whether it was projects, discussions, or collaborative learning, his insights were always sharp and well-grounded. If you’re looking for someone who combines technical depth with a collaborative mindset, I highly recommend Ahmed for any opportunity in AI/ML.
I am pleased to write this for Ahmed Islam, with whom I had the privilege of working closely during our Final Year Project (FYP). Ahmed is an exceptionally talented individual with a deep understanding of Artificial Intelligence and Machine Learning, and a genuine passion for using technology to solve real-world problems. Throughout our FYP, Ahmed consistently demonstrated remarkable dedication and technical proficiency. He was not only well-versed in core AI and ML concepts but also proactive in exploring new frameworks, optimizing models, and integrating innovative solutions into our project. His ability to troubleshoot issues and find efficient workarounds played a crucial role in the success of our work. I have no doubt that Ahmed will excel in any opportunity he pursues and will continue to contribute significantly to the field of AI and Machine Learning. I strongly recommend him without reservation.
He was amazing to work with and really compromising personality ❤
How I build.
An agent that demos well and an agent that survives production are different artefacts. Nothing here ships until it has been measured against a rubric, not a hunch.
1 · Scope
2 · Design
3 · Build
4 · Evaluate
5 · Ship
- MCPtool surface
- FAISS + ChromaDBmemory
- LangGraphorchestration
- Langfusetracing
- Docker + CIdelivery
No agent ships unevaluated
Things I’ve shipped.
10 builds3 live7 open source
11 certifications10 issuers
Certified. Then applied.
Tell me what you need.
Pick what you need, tell me what you’re building, and I’ll come back with a scope. I reply to everything within a day or two.