RAG & LLM Applications
Chatbots and knowledge assistants grounded in your documents and databases, with citations, evaluation and guardrails.
Open to remote roles · full-time, contract & freelance
Hi, I'm Muhammad Ahmad.
I build production AI that ships.
AI/ML Engineer with 4+ years designing, training and deploying RAG systems, AI agents, computer vision and generative models, from PyTorch experiments to FastAPI services in production.
- Years in production AI
- 4+
- AI projects shipped
- 15+
- ASL recognition accuracy
- 99.5%
- BS Computer Science CGPA
- 3.72
About
From research to real products
I'm an AI/ML Engineer at Arhamsoft (Pvt) Ltd in Lahore, where since 2022 I've led end-to-end delivery of AI products: object detection, OCR pipelines, speech-to-video and image-to-video generation, and retrieval-augmented generation and multi-agent systems that automate real business workflows.
I enjoy the whole lifecycle: reading the research, training and evaluating models in PyTorch or TensorFlow, then turning them into reliable, cost-efficient services with FastAPI, Docker and the cloud. I care about measurable results, clean code and systems that keep working after the demo.
- Lahore, Pakistan (UTC+5), working remotely with global teams
- Overlaps with European hours and US mornings
- BS Computer Science, UMT Lahore (CGPA 3.72)
What I build
AI solutions for teams and clients
Available for full-time remote roles, contracts and freelance projects.
AI Agents & Automation
Tool-using agents and multi-agent workflows with LangGraph, CrewAI and n8n, including human-in-the-loop approvals.
Computer Vision
Detection, tracking, counting and classification with YOLO, OpenCV and PyTorch, from webcam to CCTV streams.
Document AI & OCR
Turn invoices, receipts and scanned PDFs into validated structured data with OCR and LLM extraction.
Voice AI
Real-time voice agents with speech-to-text, LLM reasoning, tool calling and natural text-to-speech.
Generative AI & Deployment
Diffusion and VAE pipelines, MCP servers and LLM integrations, shipped as Dockerised APIs on AWS, GCP or Azure.
Portfolio
Selected projects
Open-source flagships with tests and docs, plus work from client and university projects.
Loading projects…
Experience
Where I've worked
-
Apr 2022 – Present
AI/ML Engineer
Arhamsoft (Pvt) Ltd · Lahore, Pakistan
- Lead end-to-end delivery of AI/ML projects spanning object detection, speech-to-video generation, generative modelling and OCR pipelines, from prototype to production.
- Architect and deploy retrieval-augmented generation (RAG) and multi-agent systems with LangChain, LangGraph and CrewAI to automate complex business workflows.
- Engineer image-to-video and audio-to-video generation pipelines using diffusion models and variational autoencoders (VAEs).
- Integrate trained models into scalable REST API services built on Flask, FastAPI and Django with backend and product teams.
-
2020 – 2024
BS Computer Science
University of Management and Technology · Lahore
- CGPA 3.72.
- Final year project: SignEase, American Sign Language recognition with 3 integrated AI models and 99.5%+ accuracy on alphabets and digits.
Research & development
Unstructured data parsing
Benchmarked frameworks for converting complex PDF and Excel tables into structured data; LlamaIndex proved the most robust for parsing and contextual extraction.
Sign language recognition
Compared pose-estimation datasets and deep learning architectures for production-level ASL recognition, reaching 99.5%+ accuracy on digits and alphabets.
Spectral reconstruction
Evaluated high-dimensional transformation methods and identified MST++ as the best approach for converting 3-channel RGB images into compact 31-band representations.
Toolbox
Skills & technologies
Generative AI & LLMs
Machine & Deep Learning
Computer Vision & OCR
Backend & MLOps
Languages & Data
Automation
Contact
Let's build something useful
Hiring for a remote AI role or need an AI project delivered? Send a message and I usually reply within a day.