Asad Ali

I am a results-driven Full-Stack Engineer and AI/ML Practitioner with a strong track record of building and deploying high-impact software and generative AI solutions across healthcare, fintech, trading, and digital media. With 10+ years of experience spanning backend architecture, cloud infrastructure, and large-scale ML systems, I’ve contributed to mission-critical projects — from modernizing financial products for NBFCs to fine-tuning foundation models for safe and explainable AI. As a technical author for GeeksForGeeks and Geekflare, I regularly share engineering knowledge with the developer community. I also played a pivotal engineering role in launching WOO, a dating app that scaled to 3M+ users within six months. At Collinear.ai, I led the end-to-end design and development of multiple GenAI products, integrating state-of-the-art LLMs with scalable MLOps pipelines: - Built an automated red-teaming and vulnerability assessment framework using GPT-4, LLaMA-2, and Mistral-7B to evaluate model robustness under adversarial prompts. - Developed real-time hallucination detection and content safety layer powered by RLHF, Fireworks Guardrails, AWS Lambda, and custom moderation chains. - Engineered feedback-driven prompt optimization and fine-tuned foundation models like LLaMA-2, Mistral, and Falcon using LoRA and QLoRA on healthcare and finance data. - Created a backend experimentation library for synthetic data generation, multi-model evaluation, and self-instruct fine-tuning pipelines for agentic AI copilots. Used Python (FastAPI, PyTorch, LangChain) for backend/ML pipelines, React for frontend UIs, and EKS, Terraform, and CI/CD pipelines for GPU orchestration and deployment — reducing iteration time by 40%. Previously at Johnson Matthey, I designed and implemented: A Trade Order Management System (TOMS) for managing RFQs, limit/reference orders, and forex/commodity trades with high throughput and compliance. A Metal Transfer system to manage secure asset transfers between accounts, using Java, Python, React, and Azure-native services. In the fintech sector, I: Built loan lifecycle systems for NBFCs (consumer, agent, and device loans). Integrated KYC modules and digital payments to automate onboarding. Tech Ecosystem: AI/ML: Hugging Face, Transformers, LoRA, QLoRA, RLHF, LLMOps (Bedrock, SageMaker, SkyPilot, MLflow) Cloud/Infra: AWS (Lambda, SQS, EKS), GCP, Azure, Terraform, Docker, Kubernetes, Jenkins Backend: Python, FastAPI, Java, Spring Boot, Celery, Redis, Kafka Frontend: React.js, Next.js, Redux/RTK Databases: PostgreSQL, MongoDB, MySQL