Malik Abdul Rehman

Malik Abdul Rehman – Generative AI Engineer with 3 years of actual, hands-on professional experience in freelance and project-based roles for various clients, focusing on building, fine-tuning, and deploying Large Language Models (LLMs), Retrieval-Augmented Generation (RAG) systems, and multimodal AI solutions.

From June 2022 to Present (3 years), as a Generative AI Engineer in freelance/project-based capacities, I've delivered AI services to multiple clients, including startups and enterprises. My key job roles and descriptions include:

- Designing and fine-tuning transformer-based LLMs such as GPT, BERT, T5, and LLaMA to improve response relevance and reduce hallucinations in conversational agents, using prompt engineering and optimization techniques.

- Building multimodal pipelines that integrate LLMs with diffusion models and GANs for computer vision tasks, creating interactive digital avatars with text-to-image and audio synthesis capabilities.

- Implementing RAG (Retrieval-Augmented Generation) systems for enterprise document Q&A, which involved data preprocessing, vector database integration (e.g., Pinecone), and orchestration with tools like LangChain, resulting in improved answer accuracy on domain-specific data and reducing manual effort by up to 80%.

- Deploying production-grade inference services in cloud environments (e.g., AWS, GCP) using Docker, with a focus on model optimization for low-latency serving, cost reduction through quantization, and scalable API development.

- Collaborating with product and design teams to prototype and ship user-focused AI features on schedule, ensuring seamless integration and high performance.

These experiences directly reflect my resume, where I've listed the same roles and achievements under "Professional Experience" and "Selected Projects." For example:

- Custom Chatbot with Animated Avatar: Developed an avatar-backed chatbot with context-aware dialogue, increasing simulated user engagement in demos.

- Humanlike Voice Chatbot: Built end-to-end TTS and speech UX for natural, real-time voice interactions.

- Automated Text Preprocessing Pipelines: Created LLM-based cleaning and normalization tools, reducing manual data prep time for downstream models.

My skill set includes Python (TensorFlow, PyTorch, NumPy, Pandas), transformer architectures, LLMs & prompting, RAG/document Q&A pipelines, computer vision (Diffusion, GANs), text-to-image & audio synthesis, model deployment (Docker, Cloud), data preprocessing & synthetic data, evaluation & optimization, and product-focused AI integration—all honed through 3 years of practical freelance work.

I hold a degree from the National University of Modern Languages (NUML) and certifications in Python Programming for All Levels (PSDI), Generative AI with Python (FreeCodeCamp.org), and Machine Learning Basics (PNY.pk).

Passionate about delivering truthful, high-impact AI solutions, I'm available for AI/GenAI roles. Samples of previous work uploaded in Portfolio section, including GitHub links to projects.