Open to opportunities

Muhammad Abdullah

@muhammadabdullah31

Message

Full Stack AI engineer crafting agentic LLM systems and scalable automation with measurable business impact.

Pakistan

Message

What I'm looking for

I’m looking to build production-grade agentic LLM systems—LLM orchestration, RAG, and RLHF—paired with reliable backend engineering. I value Agile teams, rigorous evaluation, and clear impact on accuracy and automation.

I’m a Full Stack AI Engineer focused on agentic AI and LLM orchestration—building autonomous workflows that turn complex objectives into reliable, production-ready automation. I bring deep expertise across RAG pipelines, fine-tuning, and RLHF, with a strong bias toward measurable outcomes.

Across healthcare, finance, marketing, and trading, I’ve delivered production-grade AI solutions that reduce manual effort by 30–50% and improve accuracy. I specialize in designing autonomous AI workflows and real-time automation platforms, backed by scalable backend architectures.

Recently, I’ve been Embedded as an AI Engineer for Google (via Turing), validating, optimizing, and debugging large-scale production AI codebases to improve correctness, performance, and reliability. I designed prompt engineering strategies and implemented prompt iteration and RLHF workflows to align model behavior with human feedback and domain-specific requirements.

Before that, at Nexxt.ai I architected and maintained secure, scalable AI-powered platforms—building a voice agent supporting 20+ concurrent calls with sub-1000ms end-to-end latency, and deploying low-latency event-driven ML for predictions. I also developed resume parsing for 1,000+ resumes, image-based product search using CNN features, and productionized ML with FastAPI and SageMaker using model versioning, retraining triggers, logging, and inference monitoring.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer

Current

Turing

Feb 2025 - Present (1 year 2 months)

Embedded as an AI Engineer for Google, validating, optimizing, and debugging production LLM workflow code to improve correctness, performance, and reliability. Designed prompt strategies and implemented prompt iteration and RLHF workflows to align model behavior with human feedback and domain requirements.

Prompt Engineering Python Rust Go Model Evaluation Production AI Systems

AI Engineer

Nexxt.ai

Jul 2023 - Jan 2025 (1 year 6 months)

Architected and maintained secure, scalable AI platforms across Fintech, E-commerce, and Real Estate, focusing on compliance, real-time processing, and resilience. Built a voice agent supporting 20+ concurrent calls under sub-1000ms end-to-end latency and productionized ML systems with FastAPI and AWS SageMaker, including automated retraining and safe CI/CD model promotion.

fastAPI AWS Sagemaker Voice AI Random Forest XGBoost CNNs (VGG16 ResNet50)Event Driven Architecture

Machine Learning Engineer

Ascendia AI

Jan 2022 - Jun 2023 (1 year 5 months)

Developed production-grade recruitment platforms combining real-time voice AI, RAG semantic search, and event-driven backends with low latency. Implemented a WebRTC/LiveKit interview system achieving sub-500ms streaming transcription latency and delivered hybrid RAG matching with 100x faster sourcing and 85%+ match accuracy.

LiveKit WebRTC LangGraph RAG Deepgram AWS Lambda ONNX Runtime

Software Engineer

Infostack Pakistan

Feb 2019 - Dec 2021 (2 years 10 months)

Built secure backend services, event-driven workflows, and data pipelines for healthcare and financial platforms. Improved API response times by ~40% using Redis caching and indexing, implemented JWT/OAuth 2.0 with RBAC, integrated Stripe with idempotency/retries, and automated compliance data collection via 12 Selenium/Scrapy scraping projects.