You will get Multilingual NLP Solutions for Low-Resource Languages
Project details
Few freelancers specialize in Urdu and low-resource languages—I do, with multiple public Urdu TTS datasets, TinyStories-Urdu, and fine-tuned models on Hugging Face. Combined with my broader LLM expertise (100+ quantized models), I deliver culturally accurate, high-quality solutions for South Asian and underserved language markets.
Specialist in Urdu and low-resource language AI solutions. I create fine-tuned models, custom datasets, and TTS systems for regional applications.
Portfolio includes multiple Urdu TTS datasets, TinyStories-Urdu, and Urdu-focused models on Hugging Face.
Specialist in Urdu and low-resource language AI solutions. I create fine-tuned models, custom datasets, and TTS systems for regional applications.
Portfolio includes multiple Urdu TTS datasets, TinyStories-Urdu, and Urdu-focused models on Hugging Face.
Machine Learning Tools
BERT, ChatGPT, fastText, Google AutoML, GPT-3, Keras, MLflow, NLTK, NumPy, NVIDIA AI Platform, Open Neural Network Exchange, OpenCV, pandas, Python, Python Scikit-Learn, PyTorch, scikit-learn, SciPy, Sonnet, Tesseract OCR, XGBoostWhat's included
| Service Tiers |
Starter
$350
|
Standard
$650
|
Advanced
$1,200
|
|---|---|---|---|
| Delivery Time | 4 days | 7 days | 12 days |
Number of Revisions | 0 | 1 | 2 |
Number of Model Variations | 1 | 1 | 2 |
Number of Graphs/Charts | 4 | 6 | 10 |
Model Validation/Testing | |||
Model Documentation | - | ||
Data Source Connectivity | - | ||
Source Code |
Optional add-ons
You can add these on the next page.
Fast Delivery
+$30 - $100
Additional Model Variation
(+ 2 Days)
+$100
Data Source Connectivity
(+ 2 Days)
+$100About Ahsan
AI/ML Engineer & Researcher | LLM Fine-Tuning, Agentic AI & Multimodal
Peshawar, Pakistan - 1:45 am local time
My open-source contributions—including QuantLLM (for 4/8-bit quantization) and huggingface-lifecycle (MLOps for HF workflows)—along with 100+ models on Hugging Face, demonstrate my commitment to practical, deployable AI. Whether you need a custom LLM, an intelligent agent workflow, or a full-stack AI integration, I deliver tailored, high-performance solutions that drive real results.
*My Expertise Includes:
*Generative AI & Large Language Models
- LLM fine-tuning with QLoRA/PEFT, RAG pipelines, and efficient training paradigms.
- Model quantization (4-bit/8-bit/GGUF) for consumer-grade hardware using tools like Bitsandbytes and my QuantLLM library.
- Building lightweight transformers (e.g., FaseehGPT for Arabic morphological understanding and fast inference).
- Integration with LangChain, CrewAI, and LangGraph for agentic and multimodal systems.
*Computer Vision & NLP
- Custom CNN architectures and medical imaging models (92% AUC in skin cancer detection).
- Real-time systems (emotion recognition with OpenCV/TensorFlow).
- Multilingual/low-resource NLP, including non-English transformer optimization.
*Full-Stack AI Development & MLOps
- Backend services with FastAPI/Flask, Streamlit/Gradio apps, and Docker deployment.
- Cloud platforms: AWS (SageMaker, Lambda), Google Cloud (Vertex AI).
- Vector databases (Pinecone, ChromaDB, FAISS) and end-to-end ML lifecycle management.
- Frameworks: PyTorch (primary), TensorFlow, Hugging Face Hub.
*Notable Projects & Open-Source Contributions
- *FaseehGPT → Lightweight transformer specialized for Arabic text generation with enhanced morphological accuracy and low-resource inference (developed at AlphaTechLogics).
- *QuantLLM → Open-source library for efficient LLM quantization, enabling massive models on consumer hardware (GitHub).
- *huggingface-lifecycle → Production-ready Python package for managing Hugging Face training workflows, versioning, and deployment.
- *DocsGPT → RAG-based documentation assistant using LangChain for intelligent query handling, summarization, and code generation.
- *torchvision-customizer → Intuitive API for building and modifying custom CNNs from scratch.
- *Medical Diagnostics Models → High-accuracy CNNs for skin cancer (92% AUC) and pneumonia detection, plus real-time emotion recognition systems.
*Experience Highlights
- *AI Engineer @ AlphaTechLogics (Feb 2025 – Present): Engineered FaseehGPT, built agentic AI systems reducing manual workflows by 50%, and led enterprise LLM fine-tuning with FastAPI integrations.
- *AI/ML Researcher @ Islamia College University (Jan 2025 – Present): Leading research in multimodal systems, mentoring students, and authoring papers on transformer optimization for non-English languages.
- *AI/ML Intern @ DevelopersHub Corporation (Sep–Dec 2024): Developed diagnostic CNNs and edge-optimized vision systems.
*Education & Certifications
- Bachelor of Science in Artificial Intelligence – Islamia College University (Oct 2023 – Present)
- Certifications: Deep Learning Specialization (DeepLearning.AI), Generative AI with LLMs (AWS), TensorFlow Developer (Google), NLP with Transformers (Hugging Face)
*What I Offer
- *Proven Results: Public open-source projects and live demos as proof of quality and expertise.
- *End-to-End Delivery: From research and prototyping to optimized deployment and documentation.
- *Clear Communication: Regular updates, transparent milestones, and client-focused revisions.
- *Efficient & Scalable Solutions: Specializing in GPU-poor and low-resource setups without sacrificing performance.
Steps for completing your project
After purchasing the project, send requirements so Ahsan can start the project.
Delivery time starts when Ahsan receives requirements from you.
Ahsan works on your project following the steps below.
Revisions may occur after the delivery date.
Target language and specific task (e.g., Urdu TTS, Punjabi translation).
Any existing data/text samples you have.