Hussnain Arshad

Hussnain Arshad

AI Engineer
Pakistan
English, Urdu
Full-time Hybrid

About Me

AI Engineer with 4+ years building and shipping production Generative AI systems: LLM chatbots, RAG pipelines, multi-agent workflows, text-to-SQL, and document intelligence. Hands-on with LangChain, LangGraph, CrewAI, MC…

Experience

Senior Machine Learning Engineer

7Vals · Lahore , Punjab , Pakistan
Apr 2025 - Present · 1 year 6 months

<p>Designed and deployed production Generative AI / LLM features across four enterprise platforms, from POC to production.</p><p> Built RAG chatbots, agentic workflows, text-to-SQL automation, AI reporting, and semantic search to automate complex enterprise workflows.</p><p> Architected LLM orchestration and multi-agent systems with LangGraph, LangChain, OpenAI, Anthropic, and MCP for task routing and enterprise API interaction.</p><p> Developed an authenticated MCP server for product APIs (asset create/retrieve/update), cutting manual operational effort by ~80%.</p><p> Automated asset creation with OCR and Whisper speech-to-text, reducing manual data entry by ~90%.</p><p> Shipped Smart Diagnosis, AI Summaries, Similar Tickets, Recommendations, and Smart Ticket Assignment to speed support workflows.</p><p> Built Python/FastAPI microservices; deployed containerized AI workloads with Docker, AWS Lambda, and ECR.</p><p> Improved AI response accuracy by ~20% and cut API latency by ~35% via evaluation, prompt engineering, semantic filtering, and caching.</p><p> Led a 3-person AI innovation team delivering AI capabilities across four core products with product and engineering partners.</p>

ML/AI Engineer

Premier NX · Lahore , Punjab , Pakistan
Feb 2024 - Mar 2025 · 1 year

<p>Deployed production LLM apps with LangChain, Mixtral, Claude, and Azure OpenAI, improving response accuracy by ~20%.</p><p> Architected hybrid RAG pipelines with PGVector and PostgreSQL, reducing manual information retrieval effort by ~60%.</p><p> Orchestrated multi-agent workflows with CrewAI, cutting manual intervention in repetitive processes by ~70%.</p><p> Built document intelligence pipelines with Azure Document Intelligence, reducing registration-card processing time by ~90%.</p><p> Integrated Azure Speech Services for speech-enabled AI experiences; shipped secure FastAPI microservices with authenticated REST APIs.</p><p> Containerized and deployed AI services on Microsoft Azure with Docker for scalable production execution.</p>

Machine Learning Engineer

One Machine Software · Lahore , Punjab , Pakistan
Jun 2022 - Jan 2024 · 1 year 7 months

<p>Built a multilingual medical AI chatbot with LangChain, Pinecone, and Whisper covering 800+ medications.</p><p> Created a domain medical knowledge base from government and clinical sources; evaluated OpenAI, Llama, Mistral, and AWS Bedrock for accuracy, latency, and cost.</p><p> Automated insurance document verification with PaddleOCR and Tesseract (~50% less manual entry); implemented text-to-SQL for natural-language healthcare queries.</p><p> Deployed containerized AI services on Azure with Docker (99.9% uptime); used CrewAI agentic workflows to cut insurance claim turnaround by ~40%.</p><p> Built NLP pipelines for sentiment analysis and intent detection; ran internal workshops on prompt engineering and LLM evaluation.</p>

Skills

Python FastAPI REST APIs TensorFlow Keras LangChain LangGraph CrewAI MCP OpenAI Anthropic Amazon Bedrock Ollama RAG text-to-SQL fine-tuning prompt engineering multi-agent systems Pinecone PGVector PostgreSQL OCR PaddleOCR Tesseract Whisper document parsing semantic search AWS Lambda ECR Microsoft Azure Docker Git
Report this Profile?