Farhan Laiq

AI Engineer at Expert System Solutions

Architected and deployed production RAG chatbot with Next.js frontend, Python backend, and Pinecone vector database, serving 1000+ daily queries with 95% accuracy using OpenAI embeddings and n8n workflow automation. Built Meta Ads Advisor chatbot with real-time campaign analytics and optimization suggestions using LangGraph agents, delivering ~30% performance improvements for client ad campaigns. Developed a law assistant chatbot for three jurisdictions with document ingestion pipeline and admin controls for knowledge base management, reducing legal research time by 60%. Engineered real-time computer vision pipelines using YOLOv11 for shuffleboard analytics, aligner defect classification (10+ categories), and drone-based smart city surveillance (parking violations, potholes, infrastructure issues). Designed end-to-end ML pipelines including dataset generation, annotation (Roboflow), preprocessing, model training and evaluation; integrated Llama3 and GPT models into intelligent applicat

AI Engineer at Hygen X

AI Engineer working on advanced AI solutions including Defective Aligner Analyzer, RealTime ShuffleBoard Analyzer, AI-Powered Invoice Text Extractor, and Liveness Detection. Implemented 1-1 and 1-N Face Verification, Intrusion Detection System using ML, and dataset generation with preprocessing pipelines. Utilized YOLO v11 embeddings, FastAPI endpoints, and Docker-based deployment; integrated OpenAI embeddings and large language models (LLMs) for contextual reasoning.

AI Engineer at ESS

Spearheaded development of breakthrough real-time detection systems using YOLOv11, creating solutions revolutionizing client monitoring across sports analytics, manufacturing quality control, and smart city infrastructure. Transformed data processing efficiency by engineering intelligent pipelines, enhancing annotation accuracy, and optimizing model evaluation processes, achieving 40% reduction in training time while maintaining high machine learning standards. Pioneered integration of state-of-the-art transformer architectures (Llama3, OpenAI models) into enterprise-grade applications leveraging advanced NLP to build truly intelligent solutions responsive to user needs with unprecedented sophistication. Fostered innovation through dynamic cross-functional collaboration, delivering scalable AI solutions exceeding client expectations and driving measurable business impact.

AI Engineer at MegaSight

Architected and delivered full-stack AI solutions transforming enterprise client approaches to complex analytical challenges, designing sophisticated deep learning models including YOLOv8 and advanced Large Language Models integrated with NLP pipelines serving over five major enterprise clients across diverse industries. Created and fine-tuned facial recognition systems delivering 95%+ accuracy for mission-critical security applications. Revolutionized compliance workflows with groundbreaking RAG-based chatbot systems intelligently processing documentation, achieving 60% reduction in manual processing time. Orchestrated end-to-end machine learning solutions encompassing data strategy, model architecture, rigorous testing, deployment, and performance monitoring ensuring exceptional reliability and scalability in demanding production environments.

Data Scientist at Freelancer

Crafted bespoke facial recognition and emotion detection systems exceeding client expectations, utilizing state-of-the-art deep learning architectures tailored to diverse client requirements and industry standards. Delivered real-time Text-to-Speech and Speech-to-Text solutions leveraging pre-trained transformer models achieving 90%+ accuracy and seamless user experiences. Successfully deployed comprehensive ML workflows using Flask and FastAPI frameworks combined with robust cloud infrastructure for scalable, reliable performance, earning 100% client satisfaction and multiple referrals. Excelled in remote collaboration and project management, building strong relationships with international clients through effective communication, cultural sensitivity, and consistent delivery of high-quality solutions.

Self Employed Programmer at Aura.pk / MegaSight

Delivered multiple ML/AI projects including production-grade chatbot enhancements, real-time analytics dashboards, and robust data pipelines. Implemented model fine-tuning, data annotation, and end-to-end deployment workflows leveraging FastAPI, Next.js, and cloud services for scalable solutions.

Self Employed Programmer at Aura.pk / MegaSight

Expanded AI chatbot capabilities and dashboards; enhanced data pipelines for scraping and ingestion; integrated OpenAI embeddings and vector stores for knowledge base improvements; implemented admin features to curate documents and ensure response accuracy; supported analytics through data retrieva and day-based insights.

Data Scientist Programmer at Force

Built and deployed real-time object detection and liveness detection models achieving 90%+ accuracy for KYC verification pipelines. Developed comprehensive web scraping tools and feature engineering pipelines supporting data-driven decision-making. Implemented credit card fraud detection system using ensemble machine learning models reducing false positives by 25% and improving detection efficiency. Utilized Transformers, Flask-SocketIO, and GitLab CI/CD to deploy scalable systems with automated testing and monitoring.

AI Engineer at Aura.pk / MegaSight

Led ML/NLP and CV projects, including dataset generation, data preprocessing pipelines, annotation, and model fine-tuning (YOLO v11), with FastAPI endpoints and model evaluation. Built real-time computer vision components (object detection, liveliness detection) and integrated with a Next.js frontend and vector database. Implemented end-to-end data ingestion and monitoring dashboards to support analytics and product decisions.

Web Developer at Aura.pk

Developed responsive web applications and gained foundational experience in full-stack development, establishing the technical foundation for ML engineering career.

Self Employed Programmer at Aura.pk / MegaSight

Developed and deployed production-level AI chatbot and dashboard for data analysis, including web scraping, data ingestion, and a vector database-backed knowledge base. Built end-to-end ML/NLP workflows with Python backend and Next.js frontend, integrated Pinecone for vector storage, and established workflows to manage chatbot responses and data ingestion. Created a country/region-aware law assistant chatbot with admin controls to add/remove documents, ensuring accurate responses. Analyzed metadata to provide actionable insights for analytics and ad performance.

AI Engineer at Programmer Force

Built KYC verification pipeline with 1-1 and 1-N face verification using FaceNet embeddings and Qdrant vector database, processing 5000+ daily verifications. Implemented liveness detection by combining YOLOv8 face detection with anti-spoofing models (98% true positive rate). Developed compliance reporting chatbot using RAG architecture with OpenAI embeddings and Qdrant, reducing manual report generation time by 70%. Delivered 10+ end-to-end AI solutions including facial recognition, emotion detection (CNNs), and speech-to-text/text-to-speech using transformers. Built cashback automation system scraping 15+ competitor sites, validating deals with OpenAI, populating Google Sheets with unique offers. Deployed scalable ML apps using Flask/FastAPI with cloud integration, ensuring high availability. Developed credit card fraud detection system with ensemble models (XGBoost, Random Forest) achieving 94% precision and 89% recall. Designed intrusion detection system for network security, reduci

Software Engineer at Turing

Analyzed datasets using Python libraries (NumPy, Pandas) to extract insights and automate tasks; contributed to training large language models (LLMs) by curating reasoning data and validating outputs; collaborated on RLHF-based LLM training, evaluation, and alignment; designed evaluation workflows for multi-turn AI systems; curated and annotated large-scale reasoning datasets for supervised fine-tuning and RL pipelines; automated data processing and model assessment using Python and ML frameworks; worked with cross-functional AI teams to improve scalability, robustness, and efficiency of LLM evaluation systems; contributed to model behavior improvements via feedback analysis, error identification, and iterative refinement.