Looking for a job

Muhammad Talha

@talhadilshad

AI Engineer at InstantAI Labs, leading applied ML initiatives across healthcare, industrial safety, and content platforms.

Pakistan

Message

What I'm looking for

I’m an AI/ML engineer with five years of experience. My work spans generative AI and LLM systems, computer vision, and classical machine learning RAG pipelines, multi-agent systems, recommendation engines, and production data pipelines. I’m looking for a team working on real applied problems, and I’m comfortable collaborating across US, UK and European timezones.

At InstantAI Labs, I lead applied ML and AI initiatives spanning computer vision, generative AI, and multi-agent systems across healthcare, industrial safety, and content platforms.

At Al Daim Solutions, I worked on end-to-end data science projects, taking data from collection and pipeline design through statistical and predictive modeling. I contributed to recommendation systems and deep learning models used in production across e-commerce, retail, and logistics.

My work includes computer vision, LLMs, recommendation engines, and real-time data pipelines. I use Python and SQL alongside cloud platforms including AWS and GCP, and containerization.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer / AI Product Engineer

Current

InstantAI Labs

Sep 2023 - Present (3 years 1 month)

Lead applied ML/AI across healthcare, industrial safety and content.
• Gas leak detection: real-time YOLOv11 system monitoring 10,000+ drums on a live line (Streamlit, Flask, MongoDB, Docker)
• Medical chatbot: fine-tuned GPT-3.5/4 (Vertex AI) and LoRA Llama 3.1
• MetaPause: ML/RAG health recommender, clinically validated
• Lumina: 7-agent LangGraph content system with MCP servers

Data Scientist

Al Daim Solutions

Mar 2021 - Aug 2023 (2 years 5 months)

• Real-time POS pipeline (CDC) into AWS Redshift: reports cut from 30–60 min to under 1 min
• Shopcast recommender (Django, TensorFlow, BigQuery, Redis), live on Shopify App Store; lifted conversion and AOV
• Visual similarity search deployed across multiple e-commerce stores
• Trash detection (UNet, SegNet, YOLOv5s) on 48,450-object dataset
• Random Forest delivery-time forecasting