Islāmābād, Pakistan
I’m Python focused software engineer passionate about building production grade machine learning and backend systems that scale. My work bridges data science and engineering, turning experimental models into reliable, high performing pipelines used in real world applications. What I do best: 🧠 Machine Learning & Generative AI: Architect full-stack AI systems combining LLMs, RAG, and fine-tuning for multilingual and multimodal applications. Experience with LangChain, LangGraph, LlamaIndex, Hugging Face, PyTorch and TensorFlow. ⚙️ Backend & Infrastructure: Design and implement microservices, ETL pipelines and data ingestion systems that handle millions of records. Skilled in FastAPI, Flask, PostgreSQL/Aurora, Kafka, Docker and AWS. 🚀 MLOps & Deployment: Build reproducible ML pipelines with CI/CD, feature stores, model monitoring and A/B rollouts for consistent production reliability. 💡 Optimization & Scale: Experienced in event-driven and async architectures, Celery and Redis Streams. Let’s connect. I’m always open to collaborating on innovative AI, ML, and backend projects.
I build large scale AI and data systems powering Sila's consumer intelligence platform combining multilingual NLP, backend engineering, and production ML. - Engineered the AI Brand Health Tracker, processing 100M+ data points across 20+ Arabic dialects using Python, FastAPI, Airflow and distributed async crawling pipelines. - Designed and deployed end-to-end NLP pipelines using Transformers, TensorFlow, PyTorch and custom tokenizers optimized for dialectal Arabic. - Built scalable ETL workflows with Airflow + Docker, maintaining pipeline reliability and handling millions of events per day. - Developed vector search & retrieval pipelines, enabling real-time brand insights and semantic similarity analysis. - Implemented production ML monitoring and automated retraining pipelines, reducing model drift incidents. - Contributed to backend microservices using FastAPI, PostgreSQL, Redis, Docker and CI/CD deployment workflows.
Developed ML-driven analytics features and backend systems for enterprise data products. - Built the Spend Categorization Engine, an NLP-based ML model with a continuous feedback loop for retraining and monitoring. - Improved ETL & preprocessing pipelines (pandas, SQLAlchemy), reducing data ingestion latency. - Integrated machine learning services into backend modules using FastAPI and Docker. - Designed secure, well-tested REST APIs supporting internal analytics dashboards.
- Developed and optimized Android features for customer-facing applications, improving load times. - Integrated REST APIs, authentication flows, offline caching, and UI components in collaboration with backend teams. - Built reusable modules and maintained clean architecture, gaining foundational experience that later supported backend and ML engineering roles.