I will fine tune llm models with lora and peft


About this gig
Are you looking to fine-tune a Large Language Model (LLM) on your custom dataset for a specific business need?
I specialize in parameter-efficient fine-tuning (PEFT) using LoRA and QLoRA, allowing you to customize powerful open-source models like Llama, Mistral, and TinyLlama without needing expensive enterprise hardware.
What I offer:
- Fine-tuning on your custom dataset (text, Q&A, domain-specific data)
- Techniques: LoRA, PEFT, 4-bit quantization
- Models: Llama 3, Mistral 7B, TinyLlama, and more
- Full evaluation with loss curves and accuracy metrics
- Delivery of the trained model + inference code
Why choose me?
- I have hands-on experience fine-tuning TinyLlama on security data using LoRA and PEFT
- I build full-stack AI applications with Python, FastAPI, and Hugging Face
- I focus on clean code, clear communication, and on-time delivery
My tech stack: Python, PyTorch, Hugging Face Transformers, PEFT, BitsAndBytes, LangChain.
Whether you need a chatbot for internal documents, a domain-specific assistant, or a custom AI model for your startup, I can help. Let's discuss your project!
Get to know Aikaksh Routela
AI ML Engineer and RAG Specialist
- FromIndia
- Member sinceSep 2026
Languages
English, Hindi, Bengali
Other AI Development Services I Offer
FAQ
What do I need to provide?
Your dataset (CSV, JSON, or text files) and a description of your goal.
How long does fine-tuning take?
Typically 5–14 days depending on dataset size and model choice.
Will I own the fine-tuned model?
Yes, you receive full ownership of the trained model and all deliverables.
What if I don't have a dataset?
I can help you prepare and format your data for fine-tuning. Contact me for a custom quote.

