The AI Engineer. That's Absar.
AI,Applied
AI/ML Engineer building LLM systems, chatbots, voice agents and automation that hold up in production.
Agentic AILangChainAutomationStart small grow big
About Me (&)
Started in pre-medical, ended up writing code that thinks. What happened in between is easier to show than explain.
- '19
The pre-medical detour
Finished FSc Pre-Medical with an A+ at 98 percent. Everyone assumed medicine. I had other plans.
- '22
Computer Science, properly
BSCS at Superior University Gold Campus. Databases, networks, software engineering — and a 3.7 CGPA to show for it.
- '24
Published research
A survey on Tor's multi-layer architecture and the dark web, published in a peer-reviewed journal.
- '24
Shipping for real people
Harvard CS50 Puzzle Day, an SQL certification, and Joint-Healer — a live store with real customers.
- '25
Teaching at Stanford's Code in Place
Section Leader for CS106A. Explaining Python to strangers made me a much better engineer.
- '25
First AI role
AI Engineer intern at Deepvision.AI, building chatbots and voice agents that had to work in real time.
- '26
AI/ML Engineer at Innoviast
LangChain and LLM systems in production, plus the unglamorous work of making data handling faster.
Selected work
Built with AI,
Nine shipped builds across LLMs, automation, vision and the web — each one solved a live problem.
01
GTM Email Automation & Lead Platform
Lead outreach, intelligent reply handling and CRM sync wired through n8n, Gmail, Groq, Sheets and Slack.
02
AI Content Generation Platform
Document-grounded content, bilingual LinkedIn drafts and image generation on React, Node, Prisma and Gemini.
03
Scanoma
CNN pipeline that analyses skin photographs and flags potential cancer risk from medical imaging inputs.
04
AI Chatbot
RAG chatbots on LangChain so answers come from your documents — not the model's imagination.
05
AI Order-Taking Agent
Conversational agent that takes orders end to end, including the messy back-and-forth real customers have.
06
Emailien
Bulk outreach automation that personalises every email so volume still reads as one-to-one.
07
Virtual Paint App
Paint driven by hand-gesture detection — no mouse, no stylus, only tracked motion.
08
eLearnify
Compact learning platform for courses and content delivery, built as a full web stack.
09
Joint-Healer
Live shop for natural joint pain oil with listings and SEO aimed at wellness buyers in Pakistan.
Capabilities
What
Production AI work — models, agents and interfaces — without the theatre. Click a card for the detail.
AI & LLM Engineering
LangChain, RAG and agentic systems on your data.
- RAG chatbots grounded in your documents
- LLM integration into existing products
- Prompt and retrieval quality tuning
- Answer evaluation before anything ships
Chatbots & Voice Agents
Real-time conversation that survives interruptions.
- Text and voice agents for live use
- Context handling across turns
- Latency-aware delivery
- Graceful failure when the model is unsure
Workflow Automation
n8n pipelines across the tools you already use.
- n8n design and deployment
- Gmail, Sheets and Slack wiring
- Automated outreach and reply handling
- CRM sync and data hygiene
Machine Learning & Deep Learning
Python ML/DL with data pipelines that stay fast.
- CNN and image analysis systems
- Model development for custom problems
- Data pipeline design and optimisation
- Production-minded evaluation loops
Web & Mobile Development
Interfaces that make the AI layer usable.
- React and Node.js applications
- PHP and SQL backends
- Android where the product needs it
- Full-stack delivery with a real UI
Have a problem
Describe the outcome you need. You get a clear read on what AI can own, what should stay human, and whether I am the right person to build it.
Impact
Signals that
Working facts from production work, research and teaching — not a review wall.
- Innoviast
AI/ML Engineer building LLM systems and automation in production.
- 9 builds
Live projects across agents, automation, vision and the web.
- 1 paper
Peer-reviewed publication — methods that survive review.
- Stanford
Code in Place section leader — CS106A, remote mentoring.
FAQ
Got any
LLM and LangChain systems, RAG chatbots, real-time voice agents, and n8n automation across Gmail, Sheets and Slack. I also ship Python ML/DL — including CNNs for imaging — and the React, Node or PHP interface around it when the product needs one.
Usually that is the point. Automation sits on your existing stack — Gmail, Sheets, Slack, CRM — with an LLM layer only where judgement helps. I would rather extend a workflow your team already knows than force a new one.
Grounding and evaluation. RAG ties answers to your documents, and I test against real questions before launch. Where a wrong answer is costly, the system should admit it does not know.
I work as an AI/ML Engineer at Innoviast and take selected side projects. Scope and timeline decide whether I can commit properly — tell me what you need and I will answer straight.
Because no two builds cost the same. An LLM layer on an existing app and a custom model with its own pipeline are different jobs. Describe the problem and you get a real estimate.
Email [email protected] with the problem, what you have tried, and the outcome you want. That is enough to say whether AI is the right tool — and it is not always.