Browse categories
Explore
Fiverr Pro
English
$
USD
I will be your ai ml engineer for python, data, nlp, llm and streamlit apps
Salman Haider
About this gig
AI/ML developer with 15+ years building and shipping production software
and web applications.
That means your model does not die in a notebook. I build it, containerize
it, test it and deploy it.
MS Artificial Intelligence @ UT Austin Texas, USA
ML & DEEP LEARNING
Classification, regression, forecasting, clustering, recommenders, anomaly
detection. scikit-learn, XGBoost, PyTorch, CNNs, U-Nets, Transformers.
NLP & LLMs
Text classification, NER, sentiment, summarization, embeddings. LoRA
fine-tuning, RAG over your documents, prompt engineering, tool-using agents.
Hugging Face, LangChain, OpenAI/Anthropic APIs.
DATA
CSV, Excel, SQL databases, scraping, ETL pipelines, EDA, dashboards.
pandas, NumPy, Plotly.
HEALTHCARE AI
MIMIC-III EHR modeling, clinical notes NLP, ICD coding, clinical RAG.
ENGINEERING & DEPLOYMENT
Docker, CI/CD pipelines, Git/GitHub, FastAPI, Streamlit, Gradio, Hugging
Face Spaces, cloud deploys. Clean repos, real tests, documented handover.
Every order ships commented code, real metrics and a plain-English write-up.
Start with the Consultation package, or MESSAGE ME your data, goal and
deadline for a free custom offer.
Get to know Salman Haider
Salman Haider
- FromUnited Arab Emirates
- Member sinceAug 2025
- Avg. response time1 hour
Languages
English
I’m Salman Haider, a GitHub & Acquia Certified Drupal Grandmaster and AI Engineer with 15+ years of experience. I specialize in building high-performance Drupal platforms and integrating AI-powered features like semantic search, auto-tagging, and chatbots. I develop Agentic AI workflows, n8n-based automations, and Azure AI solutions to streamline operations. With 250+ projects delivered, I combine technical depth with a customer-focused approach to create fast, scalable, and future-ready platforms.
FAQ
My project doesn't fit your packages. What now?
Most machine learning work is custom. Message me your data (or a sample), your goal and your deadline. I'll reply with a scoped custom offer at a fixed price, free and with no obligation.
What data formats can you work with?
CSV, Excel, JSON, SQL databases (PostgreSQL, MySQL, SQLite), images, PDFs, raw text and API feeds. Messy data across multiple tables is part of the job. I handle the joins, cleaning and ETL.
Can you build a web app around the model, not just a notebook?
Yes. Streamlit, Gradio and FastAPI are my usual stack, deployed to Hugging Face Spaces, Streamlit Cloud or your own server. You get a live link you can share with anyone.
Can you fine-tune an LLM on my own data?
Yes. LoRA and QLoRA fine-tuning on open models (Llama, Mistral, SmolLM2, Qwen) via Hugging Face, plus evaluation so you can prove the fine-tuned model actually beats the base model.
Can you build a RAG chatbot over my documents?
Yes. Retrieval augmented generation over your PDFs, docs or database: chunking, embeddings, vector search (FAISS, Chroma, Pinecone) and an LLM layer via OpenAI, Anthropic Claude or open models.
Can you deploy or dockerize a model I already have?
Yes, and this is where 15+ years of software engineering pays off. Docker containers, REST APIs with FastAPI, CI/CD pipelines on GitHub Actions, and cloud deployment. Available as a gig extra.
My model is broken or won't converge. Can you fix it?
Send the repo or notebook plus the error or metrics you're stuck on. Debugging training loops, fixing overfitting, dead loss curves and data leakage is some of my favourite work. Usually fits Basic or Standard.
What do I actually receive at the end?
Commented Python source code, the trained model weights, a runnable notebook or repo, real evaluation metrics, and a plain-English write-up explaining the approach, the results and the limitations.
Do you do academic or research work?
I'll help you implement methods, reproduce published results, debug your code and understand what's happening. I don't write or submit graded assignments on anyone's behalf.
Will my data and project stay confidential?
Yes. I'll sign an NDA on request, I don't reuse client data, and I delete everything after delivery unless you ask me to keep it for follow-up work.
