I architect and build production-ready applications powered by Large Language Models. My focus is on moving beyond off-the-shelf models to create robust, secure, scalable, and context-aware AI systems — deployed across cloud and on-premises environments.
Engineering enterprise-grade RAG beyond simple vector lookups. Intelligent chunking, hybrid search (sparse + dense vectors), multi-stage retrieval with query transformation and cross-encoder re-ranking — resulting in verifiable, low-latency systems trusted in production.
Fine-tuning open-source models (Llama 3, Qwen3, Mistral) on custom datasets. Leveraging HuggingFace ecosystem (Transformers, TRL, PEFT) with parameter-efficient methods like LoRA/QLoRA to create smaller, niche-expert models.
Engineering autonomous AI agents capable of reasoning, planning, and executing multi-step workflows. Building stateful, multi-actor systems with LangGraph for automated synthetic data generation, complex analysis, and MCP-powered automation.
Deploying AI solutions across AWS and Azure — from containerized microservices to fully managed ML platforms. End-to-end MLOps with Docker, Kubernetes, CI/CD pipelines, and real-time model monitoring.
Building synthetic data generation pipelines with 94% accuracy and zero distribution loss. PII anonymization, NeMo guardrails, and enterprise-grade data security for compliance-ready AI systems.
Designing neural architectures for CV, NLP, and tabular data. Predictive analytics, real-time inference pipelines, and publishing research at top venues including ICML.
Identified and resolved issue #518 by implementing a comprehensive quality control process for checking and updating non-working links in all .rst files, leading to a 95% reduction in broken links across the website.
Implemented a switch from radixtree_uri to radixtree_host_uri in the default HTTP router with test cases, resolving route priority confusion for users and reducing latency by 15% and improving overall performance.
ML-based web app offering career mentorship, personalized roadmaps, virtual internships, and an LLM chatbot. Drove project to Imagine Cup National Finals, won Microsoft AI Hackathon, and secured Microsoft funding.
→ View on GitHubScalable document analysis platform using RAG pipeline with Amazon Bedrock, Lambda, and S3. Implemented role-based access guardrails using NeMo for enterprise-grade data security.
Live camera emotion detection for students — triggers alerts and emails to teachers when detecting signs of distress. Improved student mental health outcomes by 30%.
→ View on GitHubI write about LLMs, cloud AI, and ML engineering on DEV.to and Medium.
Step-by-step guide to deploying your own Large Language Model on AWS EC2 — from instance selection to serving inference at scale. Published on AWS Community Builders.
→ Read on DEV.toA practical walkthrough on hosting static web applications on S3 — from bucket setup to deployment in under 10 minutes. No servers needed.
→ Read on DEV.toSubmission for the Google AI Studio Multimodal Challenge — an intelligent tool that converts complex PDF documents into structured, usable data using Gemini.
→ Read on DEV.toDeep dive into Amazon Deequ for data quality — ensuring reliable ML training data through automated validation, constraint suggestions, and anomaly detection.
→ Read on DEV.toComprehensive guide to Amazon Elastic Container Service — understanding tasks, services, clusters, and how to orchestrate containerized applications at scale.
→ Read on DEV.toA comprehensive beginner's guide covering ML fundamentals — from supervised and unsupervised learning to practical implementation steps for your first model.
→ Read on DEV.toTechnical articles on Git workflows, cloud development, and AI engineering. Including guides on GitHub Codespaces and developer productivity.
→ Read on MediumVideo tutorials and talks on machine learning, cloud computing, and software engineering. Subscribe for deep-dive technical content.
→ Watch on YouTubeAchieved highest Gold rank. Mentored 10,000+ professionals and students. Built a community of 10K students, organized 2 hackathons, inter-city coding competition, and 50+ tech events. Won Microsoft Mentor award.
Led a 30-member diverse team. Secured $40K in annual resources from DataCamp and 100+ LinkedIn vouchers from Microsoft through strategic collaboration with global communities.
Supported thousands of students through mentorship and resources. Empowered women and underrepresented minorities via tech initiatives. Delivered talks at 10+ international conferences.
Technical thought leader sharing knowledge about cloud-native ML infrastructure with access to AWS resources, mentorship, and networking opportunities.
Founded Cloud Native Sahiwal. Organizing regular meetups on containerization, microservices, Kubernetes, and cloud-native technologies.
Delivered talks at international conferences representing GitHub, Microsoft, and Google. Advocating tech innovation and community building across borders.