01 / Flagship EngineeringShipped
PDF RAG Chatbot
Secure async document intelligence with hybrid retrieval, OCR, abstention, and grounded citations.
Saifullah Suleman
Applied AI Engineer
I build AI systems and dependable backends.
Applied AI engineer focused on computer vision, RAG systems, and Python backend development. Building evaluated AI prototypes, a live retail storefront, and CNC monitoring research at NUST PNEC.
Available for focused work
Available for full-time AI/ML roles in Karachi (onsite or hybrid) or remotely.
01 / Selected engineering
Systems built around reliability, evidence, and real constraints.
RAG systems, evaluated ML APIs, and backend infrastructure, followed by live product work and computer-vision research.
02 / Flagship EngineeringShipped
LLM API Gateway
Provider-neutral LLM infrastructure with stable aliases, retries, fallback chains, usage accounting, and analytics.
03 / Flagship EngineeringShipped
Lead Scoring ML API
An evaluated supervised-ML inference service with calibration, threshold policy, model integrity checks, and reproducible synthetic benchmarks.
04 / Flagship EngineeringShipped
AI Engineering Starter Kit
A reusable FastAPI foundation for production-minded AI services with typed configuration, persistence, testing, security guardrails, and CI.
Production, product & research
05 / Production & Product EngineeringProduction
Wahab Mobiles
A live React/TypeScript retail catalogue with a Fastify/PostgreSQL backend, inventory and admin tools, and ordering workflows prepared for an upcoming launch.
06 / Production & Product EngineeringPrivate Alpha · Coming Soon
LucidFence
A private, self-hosted household network-protection system for DNS filtering, safe-search enforcement, policy controls, and browser-assisted protection.
07 / Production & Product EngineeringPrivate Development · Coming Soon
Atlas
Procurement intelligence for mapping merchant SKUs to credible upstream sourcing alternatives and comparing evidence-backed acquisition scenarios.
08 / ResearchResearch Prototype
MS-ADA Intelligent Traffic Surveillance
A CPU-first academic computer-vision prototype for vehicle tracking, ALPR, accident reasoning, hit-and-run monitoring, and evidence capture.
02 / Experience
Research, software, and data experience.
Professional experience
September 2026 - Present
Research Analyst
PNEC, NUST
Contributing to Pakistan Science Foundation-funded CNC process-health monitoring research at NUST PNEC, Karachi.
- Curate CNC machining datasets through preprocessing, operation-wise segmentation, and labeling of continuous production-cycle signals.
- Develop and evaluate PyTorch/TensorFlow models for CNC process-health monitoring.
- Build vibration-signal inference components for integration with the project's IoT and edge architecture.
August 2025 - August 2026
Full Stack Developer
Wisdom Tech
Backend-focused development for business and e-commerce applications. Karachi / Remote.
- Built Python and Node.js/TypeScript backends and APIs with PostgreSQL integration.
- Contributed to Shopify implementations, testing, and deployment.
November 2024 - January 2025
Data Analysis Intern
NCBC, NED University
Data analysis internship in Karachi, Pakistan.
- Cleaned and analyzed datasets using SQL.
- Built Power BI and Tableau dashboards to communicate trends and performance metrics.
Project-based
Freelance Developer & Data Analyst
Independent
Delivered freelance work across data analysis and web development, translating client requirements into practical analyses, reports, and software deliverables. Specific client names and commercially sensitive details remain private unless explicitly approved for publication.
Leadership & community
2025–2026
President
TE Links — NED University
Led the departmental student body through a year spanning 10+ events, 6 technical sessions, 10+ partnerships, and 3 community initiatives across technical learning, professional development, student community, and university programming.
03 / Education
From telecommunications and computer vision toward applied AI systems.
September 2026 – Present
M.S. Artificial Intelligence
National University of Sciences & Technology (NUST)
Karachi, Pakistan
M.S. Artificial Intelligence at Pakistan Navy Engineering College (PNEC), Karachi. Expected graduation: 2028.
2022 - June 2026
B.E. Telecommunications Engineering
NED University of Engineering & Technology
Karachi, Pakistan
Engineering foundation spanning telecommunications and computing, culminating in the MS-ADA computer-vision Final Year Project.
04 / Research
Research work with a clear publication path.
All research →Work in progress is labeled honestly; published material links to the paper, preprint, or DOI.
Ongoing research
CNC Process-Health Monitoring
Pakistan Science Foundation-funded research on CNC machining signals: dataset preprocessing, operation-wise segmentation and labeling, model evaluation with PyTorch/TensorFlow, and vibration-signal inference for IoT and edge integration. Contributing as Research Analyst since September 2026.
Under peer review
MS-ADA: AI-Powered Intelligent Traffic Surveillance for Real-Time Accident and Hit-and-Run Detection & Tracking
Academic computer-vision research from the NED Final Year Project (June 2025 - August 2026), combining YOLO11 detection, ByteTrack tracking, EasyOCR-based Pakistani ALPR, and multi-signal accident reasoning. Tested on multiple raw traffic videos with snapshots, evidence clips, and structured event outputs. Manuscript under peer review; no publication or acceptance is claimed.
05 / Writing
Notes on AI research and engineering.
Practical perspectives on neural networks, reasoning, scientific machine learning, and the systems around them.
Neural Networks
Modern Neural Networks: Beyond the Architecture
From ANNs and attention to state-space models and hardware-aware kernels, useful progress comes from matching the model to the problem.
Read note →
AI Reasoning
Neuro-Symbolic AI: Learning With Explicit Reasoning
Neural models are strong at learning patterns. Symbolic systems make rules and deductions explicit. Combining them requires a careful contract.
Read note →
Physical AI
Physical AI: From Perception to Closed-Loop Action
Robotics foundation models and simulation are advancing quickly. Reliable physical AI still depends on sensing, control, evaluation, and recovery working together.
Read note →06 / Profile
Applied AI depth. Full-stack breadth. Backend discipline.
I build and evaluate AI systems with the backend infrastructure they need. My work spans computer vision, retrieval-augmented generation, Python APIs, and full-stack applications. At NUST PNEC, I contribute to Pakistan Science Foundation-funded research on CNC process-health monitoring, from machining datasets to vibration-signal inference for IoT and edge integration.
Availability
Available for full-time AI/ML roles in Karachi (onsite or hybrid) or remotely.
Applied AI / RAG / LLM systemsFastAPI / APIs / data systemsFull-stack product engineeringTesting / security / deployment
Verified proof
The signals behind the work.
Skills
AI & machine learning
PythonPyTorchTensorFlowscikit-learnModel training & evaluation
Computer vision
YOLO11OpenCVByteTrackEasyOCRDetectionTrackingALPR
LLM systems
RAGEmbeddingsQdrantOpenAI / Gemini APIs
Backend & web
FastAPINode.jsTypeScriptFastifyReactPostgreSQL
Engineering
RedisCeleryDockerpytestGitGitHub Actions
07 / Contact
Have a hard systems problem or a product worth shipping?
For engineering opportunities, collaboration, research, or product work, send the problem and context.