8+ years shipping production AI — from demand forecasting pipelines processing 50M+ records at an AWS Premier Partner to voice agents handling 10M+ concurrent calls at my own company. I build ML systems that move business metrics.

I'm an ML Architect & Engineer, Founder/CEO of Reallytics.ai, and AI Educator with 8+ years of progressive experience building production AI systems.

My path: Data Analyst (IoT) → Software Engineer (Afiniti, telecom AI) → Senior Data Scientist (9T5/Looper Insights, crypto + CV) → Principal ML Lead at NorthBay Solutions (AWS Premier Consulting Partner — 3 years working directly with AWS internal teams on enterprise ML) → Founded Reallytics.ai to build and ship AI products for clients across 4 continents.

I've worked as both an individual contributor shipping models and a technical leader managing cross-functional teams across 6 countries. At NorthBay I handled pre-sales and delivery for enterprise AI engagements. At Reallytics.ai I'm the technical architect behind every product, from concept to production.

MS Data Science (NUCES FAST) · BS Computer Science (COMSATS)

Founder & CEO / Technical Architect · Oct 2020 – Present

I founded and scaled Reallytics.ai — a full-stack AI company delivering production-grade GenAI, Voice AI, and Data Science systems for startups and enterprises across North America, Europe, the Middle East, and Australia. I'm the technical architect behind every product, from concept to production.

We don't just prototype — we ship intelligence that works in real environments.

Product	What It Does	Live
Novalytics	Context-aware voice assistants with sub-300ms latency and memory-persistent dialogue. Deployed across healthcare, finance, hospitality. Scaled to 10M+ concurrent calls
LLM Chatbots	RAG-powered chatbots for Finance, Retail, Healthcare, and HR using LangChain, FAISS, ChromaDB, OpenAI, Claude, and Mistral
AI Agents Platform	Agentic workflows: insurance underwriting, multilingual caregiving (100+ languages), admissions automation (50% faster). Connected LLMs to databases, APIs, and Slack
Data Analytics	Multi-platform BI hub (Power BI + Copilot, Tableau, QuickSight) with AI-powered insights and real-time streaming
Fraud Detection	Real-time ML fraud scoring for financial services
Real Talent	Platform to vet, train, and connect Pakistani AI talent to international clients

Technical depth: Fine-tuned LLaMA-2/Mistral using LoRA, QLoRA, PEFT via Hugging Face (70% API cost reduction). Served models via VLLM/CUDA. Cloud infra on AWS: SageMaker, Bedrock, Lambda, Glue, Redshift. Full CI/CD with Docker and CodeCatalyst. Led 10+ engineers across 6 countries, 20+ global client engagements.

Growing up in Lahore, I saw brilliant engineers with world-class skills but zero access to global opportunities. When I broke through to US, UK, and European clients, I promised I'd build a bridge back. That's why I created Real Talent — a platform where I vet, train, and connect Pakistani AI talent to international clients, handle contracts, and manage delivery end-to-end. Today we have a growing pool actively working globally.

Building Reallytics.ai was about proving world-class AI can come from Pakistan. Real Talent is about making sure I'm not the exception — I'm the beginning.

Over 8 years, I've consulted, built, and shipped production AI systems for these organizations — handling everything from pre-sales and architecture to hands-on delivery and team management.

MARS Inc	IBM	Cloud Kinetics	DataArt	AWS Startups	Silvertree Brands
4G Capital	Looper Insights UK	Tower Loan	Ashcroft	Verticiti	CXEX

Feb 2023 – Feb 2025 · Andover, MA (AWS Premier Consulting Partner)

Worked directly with AWS internal teams for 3 years, handling pre-sales and delivery for enterprise ML engagements.

• Demand forecasting — PySpark on EMR, 50M+ banking/retail records, 40% accuracy improvement
• AWS ML infrastructure — SageMaker, Lambda, Glue, Redshift serving 3+ clients simultaneously
• NLP document intelligence — AWS Comprehend, Rekognition, custom Transformers
• Computer vision — OpenCV + custom CNNs on Lambda, sub-200ms inference
• RAG systems — LangChain + FAISS + ChromaDB for enterprise knowledge retrieval
• LLM fine-tuning — LLaMA/Mistral with LoRA for domain-specific use cases
• MLOps — CI/CD via Docker + CodeCatalyst, ECS/ECR, automated retraining
• Led 5 data scientists across US, UK, Pakistan

Apr 2022 – Jan 2023 · Australia (Client: Looper Insights UK)

• Time-series forecasting (LSTM, Prophet) for crypto — 25% directional accuracy improvement
• Fine-tuned BERT for sentiment analysis, served at scale with VLLM
• Computer vision on Lambda processing 10K+ daily inputs
• COVID face mask detection (95%+ accuracy)
• RAG knowledge retrieval with LangChain + ChromaDB

Aug 2020 – Apr 2022 · Lahore

• ML pipelines for production call-routing optimization on one of the world's largest applied AI platforms
• ETL workflows (PySpark, SQL) reducing pipeline failures by 30%
• Feature engineering processing millions of daily call records

Jan 2018 – Jul 2020 · Lahore

• IoT sensor data analysis across 1000+ endpoints
• Anomaly detection and predictive maintenance using classical ML
• Power BI dashboards replacing manual reporting

Beyond building, I teach. I design and deliver workshops that help experienced developers level up with Generative AI and modern AI tooling.