Huzaifa
Imran
Senior Full Stack Developer · Co-Founder of Lordevs · Medical AI Researcher · Generative AI & Deep Learning Engineer
Career
Work Experience
- Led full stack delivery for scalable AI products, shipping React/Django solutions used by enterprise clients.
- Built backend APIs, orchestration pipelines, and front-end dashboards for knowledge-driven applications.
- Implemented production-ready CI/CD, containerized services, and automated deployments for fast iteration.
- Collaborated with designers and data teams to turn research prototypes into user-facing products.
- Mentored developers on code quality, architecture, and full stack best practices.
- Developed full stack AI workflows, integrating backend services with interactive front-end applications.
- Built multi-modal RAG/KAG pipelines and fine-tuned open-source LLMs for document automation solutions.
- Designed scalable APIs and database schemas to support real-time data ingestion and retrieval.
- Worked across requirements, architecture, and delivery to ship products end to end.
- Improved codebases through refactoring, testing, and automation for faster deployments.
- Built and maintained full stack applications, combining back-end services with user-facing interfaces.
- Deployed ML inference systems and supported production workflows on Google Kubernetes Engine.
- Implemented Django-based solutions and integrated serverless deployment pipelines on Cerebrium.
- Developed data-driven features for research and generative AI workloads.
- Collaborated with cross-functional teams to deliver stable, scalable applications.
Portfolio
Featured Projects
Agentic RAG platform delivering SRA & LAA compliant legal guidance with source citations — zero hallucinations.
Centralized intranet for 100+ employees across UK branches — staff directory, onboarding workflows, knowledge base, and branch-level access controls.
Multi-vendor pet marketplace with vendor onboarding, product listings, secure checkout, and order tracking — concept to live product.
Research
Publications
Automated EARS-Based Requirements Generation with Lightweight Large Language Models
We propose a novel framework for fine-tuning lightweight LLMs that can automatically rewrite software requirements from scratch. We designed and implemented a pipeline that transforms raw requirements into EARS format by fine-tuning four LLMs — bridging NLP and requirements engineering.
DermNet: Integrative CNN-ViT Architecture for Bias Mitigation in Dermatological Diagnostics
A method for reducing diagnostic bias in skin disease identification for people of color. Employs a zero-shot unsupervised lesion segmentation approach, feeding outputs into DermNet — a hybrid Vision Transformer and CNN classifier designed for equitable dermatology AI.
Academic
Education
Credentials
Licenses & Certifications
Expertise
Skills
Open to research collaborations, consulting engagements, and strategic conversations in AI, banking technology, and product.
buntyhuzaifa3@gmail.com