Waqar Ahmed
Senior Full Stack AI Engineer & AI Researcher with 4+ years of experience building AI-powered products and agentic workflows, grounded in systems thinking and customer needs.
Explore my work
Choose a project to learn more.
PashtoGPT
Research tooling for adapting language models to Pashto. Data preparation, QLoRA experiments, and evaluation; no released trained weights or measured quality gains yet.
Activity
Interested in working together? Let's build something useful.
About
Senior Full Stack AI Engineer and AI Researcher with 4+ years of experience turning customer needs into AI-powered products and secure full-stack systems. I have read 120+ research papers to deepen my understanding of AI and inform engineering decisions. My approach connects a detailed understanding of system architecture with customer workflows, product goals, and operational constraints. Work spans agentic HR automation with human approval checkpoints, SaaS products, and real-time video calling scaled to 10K daily users. PashtoGPT brings a research focus to low-resource language AI through Unicode-safe data preparation, Qwen3 instruction-tuning experiments, and held-out evaluation. My engineering focus is retrieval-augmented generation, tool-calling workflows, context engineering, and human oversight, with reliability, latency, and cost treated as product requirements. I connect React/Next.js interfaces, Node/NestJS APIs, and PostgreSQL/MongoDB data systems using TypeScript, Python, OpenAI API, and LangChain.
Key Highlights
Research
Read 120+ research papers to deepen AI understanding and inform practical engineering decisions
Systems & Product
Connect system architecture, customer workflows, and product goals to practical engineering decisions
Experience
4+ years across AI engineering, agentic workflows, full-stack products, and secure SaaS systems
Teaching
Mentored 100+ learners on AI, web, DevOps, and MVP building
Community
Organized or contributed to 5+ hackathons and student innovation events
Certifications
7 Coursera credentials across ML, AWS, DevOps, and data
Hackathons
Built and mentored teams through MVP development and launch preparation
Open Source
PashtoGPT: open-source Pashto language AI experiments, data auditing, and evaluation tooling
Innovation
Best Project Poster Award for EduAI at Tech-Talks and Posters Symposium 2025
Startups
Translated founder requirements into user flows, data models, API contracts, and MVP plans
Work Experience
GeekinateActiveSenior AI Engineer — Agentic AI & Forward Deployed Engineering (FDE)Aug 2025 — PresentAug 2025 — Present
Senior AI engineering focused on agentic HR automation and forward deployed product delivery. Translating client requirements into integrated AI-assisted workflows, secure SaaS modules, and web/mobile experiences, with human approval checkpoints for HR operations.
- Contributed to a stealth agentic HR automation project spanning recruitment and employee operations, with HR approval checkpoints.
- Owned full-stack web and mobile delivery across Space Manager, Shifa Foundation USA, Twinhub/coworking websites, Lilia web, and Lilia React Native app.
- Built Space Manager serverless SaaS modules for spaces, bookings, customers, dashboards, reports, and admin operations.
- Delivered React/Next.js pages, React Native screens, reusable UI components, backend integrations, and API-driven dashboard/admin workflows.
- Collaborated with marketing on the Shifa Foundation USA revamp, translating content and brand requirements into responsive, launch-ready website sections.
Viral Mobitech Private LimitedPastFull Stack AI EngineerDec 2024 — Dec 2025Dec 2024 — Dec 2025
Project-based full-stack work across real-time communication, booking/video workflows, security-sensitive APIs, and deployment-ready product features.
- Built full-stack features across Bondvia, Meetany, Secure Pulse, and Next-Chat using React.js, Node.js/Express, PostgreSQL, MongoDB, and Prisma ORM.
- Scaled Bondvia random video calling to 10K daily users by indexing database queries, optimizing APIs, debugging crash sources, and strengthening secure call flows.
- Improved Meetany reliability and user safety by adding security layers, safer call controls, content-safety/moderation paths, and deployment scaling support.
- Built Next-Chat OTP verification, authentication, onboarding APIs, and protected Secure Pulse product workflows.
- Supported CI/CD and observability with Docker, GitHub Actions, GitLab CI/CD, ArgoCD, Helm, Prometheus, and Grafana.
National Incubation Center (NIC), KohatPastAI Engineer / Startup Technical MentorJan 2023 — Dec 2024Jan 2023 — Dec 2024
Built startup MVPs, dashboards, backend APIs, internal tools, and product validation paths while mentoring early-stage teams.
- Built startup MVPs, dashboards, backend APIs, and internal tools using React.js, Next.js, Firebase, GraphQL, PostgreSQL, Express.js, and Node.js.
- Translated unclear founder requirements into user flows, product screens, data models, API contracts, and launch-ready MVP plans.
- Implemented JWT/OAuth authentication, containerized deployments, AWS, Kubernetes, CI/CD workflows, and product validation paths.
- Mentored early-stage teams through ideation, MVP development, launch preparation, and hackathon execution.
KUST Incubation Center (KIC), PakistanPastAI Engineer — Research ContributorJun 2023 — Jun 2025Jun 2023 — Jun 2025
Contributed part-time to full-stack and AI product direction for student founders, AI/ML ideas, and product delivery roadmaps.
- Helped teams turn AI/ML ideas into feasible web app MVP plans, full-stack architecture directions, and product delivery roadmaps.
- Guided practical model/API integration using HuggingFace, TensorFlow, MLflow, Docker, GitOps, and product-focused AI delivery practices.
Education
Kohat University of Science and Technology, Kohat, Pakistan
Bachelor of Science in Computer Science, CGPA: 3.5/4.0
Focused on software engineering, backend systems, artificial intelligence, distributed systems, and applied product development.
Projects
Open-source research toolkit for adapting language models to Pashto: source auditing, Unicode-safe data preparation, Qwen3-1.7B QLoRA experiments, Kaggle training workflows, and held-out evaluation with Belebele and native-speaker review. Includes retrieval and verifier diagnostics. Research stage: no released trained weights or measured quality gains yet.
Contributed to a stealth HR automation project at Geekinate spanning recruitment and employee operations, with HR approval checkpoints. Focused on connecting AI-assisted workflows to practical product delivery while keeping people involved in operational decisions.
Scaled random video calling to 10K daily users by improving matching reliability, database indexes, API performance, crash resistance, and safer call-session handling.
Strengthened high-traffic booking/video workflows with deployment scaling support, additional security layers, abuse-risk controls, and content-safety handling for a 10M+ user platform.
Built serverless coworking SaaS workflows for spaces, bookings, customers, dashboards, reports, and admin management.
Built backend APIs for OTP verification, authentication, onboarding, and deployment-ready communication workflows for international users.
Built an AI learning platform with course creation, adaptive quizzes, role-based access, recommendation direction, summarization/Q&A, and deployment architecture.
Built a recommendation model that turns learner progress signals into relevant content suggestions and adaptive learning support.
Built and improved company website sections focused on service positioning, client trust, responsive UX, and maintainable delivery.
Improved a coworking website experience to clarify offers, support customer decisions, and strengthen business credibility.
Improved nonprofit web flows around campaign clarity, donor trust, mobile responsiveness, and donation-path friction.
Stabilized backend APIs by resolving reliability issues, reducing delivery blockers, and improving maintainability.
Delivered web and mobile learning product work across reusable UI, app screens, backend integrations, and organization-ready workflows.
Built a business website for construction and environmental services focused on trust, service clarity, leads, and SEO readiness.
Built an AI-enabled commerce concept that improves product discovery, relevance, and buyer experience with personalization workflows.
Designed a startup operations dashboard that turns scattered founder updates into trackable metrics, records, and reporting insight.
Developer brand platform built to communicate engineering judgment, project ownership, and proof of work.
ML operations workflow exploring how models move from experiments into reproducible, deployable systems.
Messaging system prototype solving live communication with sockets, authentication foundations, and state flow.
Serverless productivity app using managed backend patterns to reduce operations and improve delivery speed.
Machine learning project focused on understanding model behavior, training loops, and image classification.
Blog
Why AI Agents Are Harder Than Chatbots
A practical article explaining why agents are harder than chatbots: tool calling, retrieval, planning, memory, workflow execution, evaluation, cost...
Real Full Stack AI Is More Than Prompting OpenAI
A complete article explaining why production AI products need UI, backend orchestration, databases, auth, vector search, tool calling, cost control...
Skills
Certifications
Stanford University & DeepLearning.AI
Mastered foundational AI/ML concepts, including supervised/unsupervised learning, regression, and neural networks using Python and TensorFlow.
DeepLearning.AI
Developed expertise in neural networks, CNNs, RNNs, and NLP/computer vision applications using PyTorch and TensorFlow.
DeepLearning.AI
Learned to deploy and manage ML models in production using AWS SageMaker, Docker, and CI/CD pipelines.
Amazon Web Services
Gained skills in AWS architecture, DevOps, and containerization with EC2, S3, IAM, and infrastructure templates.
Amazon Web Services
Explored AI/ML foundations and AWS services like SageMaker for machine learning workflows.
Amazon Web Services
Mastered DevOps practices, infrastructure as code, AI integration, AWS Lambda, and CI/CD.
DeepLearning.AI & Amazon Web Services
Built data pipelines and architectures for ML and analytics using AWS and open-source tools.
Achievements
Institute of Computing, Kohat University of Science and Technology | 2025
Received the Best Project Poster award for EduAI, an AI-powered personalized learning platform, recognized for innovation, clarity, and real-world impact at the Tech-Talks & Posters Symposium 2025. Pitched directly to Vice Chancellor Prof. Dr. Syed Zafar Ilyas, in collaboration with Hammad Ahmad and Faisal Minhas.
Mentorship and Hackathon Leadership
NIC Kohat, KUST, and student innovation communities | 2024 - Present
Mentored 100+ students and developers, and organized or contributed to 5+ hackathons and student innovation events.
Volunteer Experience
National Incubation Center, Kohat
Technical Mentor & Hackathon Organizer
Mentored startups in web development and DevOps practices, and organized multiple hackathons to foster innovation and coding skills among youth in South Khyber Pakhtunkhwa.
- Led mentorship for startups in building MVPs using React, Node.js, and AWS.
- Organized KIC Hackathon 2024, engaging 50+ participants in AI and web development challenges.
- Co-organized South KP CodeFest 2024, focusing on real-world problem-solving with 30+ teams.
- Facilitated NIC Kohat AI Hackathon 2024, mentoring participants in ML model development.
- Organized KUST Innovation Hackathon 2025, promoting DevOps, web apps, and AI solutions.
- Supported KIC CodeJam 2025, guiding 40+ students in competitive programming and full-stack development.
Kohat University of Science and Technology
Seminar Organizer & Speaker
Organized and delivered seminars on Artificial Intelligence to educate students on its applications and impact, tailored for Medical Lab Technology students and matric-level students.
- Conducted a seminar on 'AI and Its Impact in the Medical Field' for MLT students, highlighting AI's role in diagnostics, pathology, and personalized medicine.
- Led a seminar on 'AI and Its Use Cases' for matric and other students, introducing AI applications in education, healthcare, and local challenges in Pakistan.
- Demonstrated AI tools like Google's Teachable Machine to engage students in hands-on AI exploration.
- Encouraged student participation in tech events and provided guidance on free AI learning resources.