I build production-grade AI systems and conduct applied ML research, spanning multi-agent LLM pipelines, generative models, and network security. I currently architect agentic AI infrastructure at Synthapse and teach applied ML to 100+ students at iCodeGuru, where I design curriculum, run live sessions, and mentor student research projects.

I am actively seeking a funded PhD position with a Research Assistantship (RA) or Teaching Assistantship (TA) in AI, Machine Learning, or related areas for Fall 2027 at US universities. My work combines engineering depth with research curiosity. I build systems to answer open questions, not just to ship products, and I am working to formalize that research output into peer-reviewable form.

• Multi-agent coordination and emergent behavior in LLM systems
• Generative and probabilistic models for structured prediction and inverse design
• ML for network security, adversarial traffic analysis, and privacy-preserving protocols
• AI for smart infrastructure and cyber-physical systems
• On-device and edge AI for real-time inference

💬 If you are a professor working in these areas and looking for a motivated RA/TA, I would love to connect.

🏆 1st Place — UN/IYDO Hack for Earth 2026 (unDavos)

Decision intelligence platform for sustainable construction. Parses ICE v3.0 embodied carbon data and uses Groq-accelerated Llama 3.3 70B to generate explainable material-swap recommendations with live cost-vs-carbon tradeoff sliders. Addresses the construction industry's 39% share of global carbon emissions through AI-first pre-construction workflow integration.

ML Security Research · XGBoost · SHAP Interpretability · arXiv preprint in preparation

Built an ML pipeline to fingerprint and classify DNS-over-QUIC network traffic, a privacy protocol increasingly adopted to evade detection. Achieved 90–95% classification accuracy using XGBoost, Random Forest, and Extra Trees ensemble methods on real packet captures. Applied SHAP-based interpretability to surface the most discriminative packet-level features, making model decisions fully auditable for security analysis.

This project reflects my research approach: identify a real open problem, collect ground-truth data, build a rigorous ML pipeline, and make the results interpretable.

Synthapse R&D · Deployed on Google Cloud

AI-driven system for decentralized network and energy optimization across macro, medium, and micro scales. Autonomous agents communicate peer-to-peer: PSO agents optimize WiFi tower placement, ACO agents find lowest-latency routing paths, and a GNN coordinates real-time load balancing. Node behavior validated via ns3 simulations.

AI-Assisted Research Discovery & Synthesis

Agentic pipeline built to assist with literature discovery, synthesis, and research workflow automation, applying the same multi-agent design principles used in production systems to an academic research use case.

Autonomous Personal Assistant · LangGraph + Gemini 2.5 Flash · 200+ Users

Autonomous WhatsApp agent with Gmail and Google Calendar integration. LangGraph manages multi-step agent state, Gemini 2.5 Flash handles reasoning, and a Node.js bridge connects to a Python FastAPI backend for tool execution.

Production SaaS · 200+ Active Users · Zero Critical Incidents

AI influencer-brand matching platform. Containerized microservices with full CI/CD automation, AWS S3 for media storage, and AWS ECR for container registry.

• Hackathon Judge & Mentor — HEC, UET Lahore, iCodeGuru events
• Judge / Mentor — Pak Angels (startup and innovation mentorship)
• Active participant in 16+ international hackathons, with consistent placements across AI, sustainability, and infrastructure tracks
• Applicant, International Volunteer Forum 4.0, Budapest
• Mentor applicant, Hack for the Futures, London Climate Action Week

I treat teaching as a research activity. At iCodeGuru, I:

• Design and deliver ML curriculum to 100+ students across cohorts
• Mentor students through end-to-end research and project development
• Run structured office hours, evaluate project work, and give structured feedback

This experience directly maps to Teaching Assistantship responsibilities in a PhD program. I am not new to the classroom.

VP of AI Systems — Synthapse, Poland (Remote)  ·  2024–Present

Architect production multi-agent LLM pipelines at scale, leading system design for agentic AI infrastructure using LangGraph and LangChain. Translate open-ended technical problems into deployed, evaluated systems, the same iterative problem-to-solution cycle central to research work.

AI Full-Stack Engineer — Ingenious Zone, Faisalabad  ·  2023–Present

End-to-end AI product development: model integration, API design, and cloud deployment.

AI Instructor & Mentor — iCodeGuru, USA (Remote)  ·  2023–Present

ML curriculum design, live instruction, and research mentorship for 100+ students.