Usama Shahid

LLMs & multimodal models · RAG & retrieval · Agents & multi-agent systems · Deep agents · LangChain / LangGraph · Prompt engineering · Fine-tuning & evals · LLM orchestration · Vector databases · Open-source models · Structured outputs & tool use

Machine learning · Deep learning · NLP · Statistical modelling · Causal inference · Hypothesis testing · Text mining · Geospatial & network analysis · Trading & quantitative algorithms · Visualisation

Python / FastAPI · R · C++ · Javascript · APIs & microservices · Docker & containers · AWS (S3 · ECS · ECR) · Azure · GCP · Distributed systems (PySpark) · Databricks · MLOps & AI lifecycle · Weights & Biases · SQL · NoSQL · vector DBs · Vibe coding

Curriculum design · Technical workshops · AI literacy programmes · 1:1 coaching & mentoring · Research supervision · Stakeholder engagement · Team leadership · Public speaking

AI governance & policy · Bias & fairness auditing · Explainability (SHAP · LIME) · Safety evals & red-teaming · Hallucination mitigation · Human-in-the-loop design · Privacy & GDPR · Model monitoring & drift · Responsible GenAI adoption

Workshops across the AI/ML lifecycle, from concept to deployment and monitoring, plus one-to-one coaching. Authored the commercial AI Accelerator for NISRA (UK Government) and Citi's Level 6 AI Engineer programme.

Taught degree-level AI/ML modules, up-skilling professionals from Rolls-Royce, Mastercard, Visa, BBC, Barclays and more.

AI and data science consultancy for startups and enterprises — building systems meant to run in production, not just demo well.

• Subconscious.ai — LLM-based bio-equivalent survey respondents; replicated 200+ research studies at ~0.7 correlation.
• Girard Sudron — AI data engineering & automation — major savings in operational time and cost.
• Strategically AI — Search-engine analysis system feeding AI content generation — 70+ SurferSEO scores.
• Bramework — AI models for content generation — drafts, paraphrasing and grammar with LLMs.

Steering potential research publications in AI & data science through the peer-review process.

Led Level 4–7 modules for MSc and BSc students — Python, data analysis, ML, computational mathematics and research methods — and supervised postgraduate research in applied AI.

Built and led a 15-person engineering and AI team. Architected a GenAI system — while OpenAI's models were still in beta — automating lead generation, enriching customer data and personalising communication at scale. Employee of the Year.

• Numin — led development of data-driven algorithms for quantitative and algorithmic trading.
• RedCloud — led ads & marketing data analysis and integrated statistical data channels into the dashboard; RedCloud went on to list on NASDAQ.
• AEO — led the data science & warehouse design team, turning raw statistics into numbers the business could act on.

Developed and shipped web & mobile apps for SimSim Markets, owning app-store submissions and releases.

University of Gloucestershire · 2025 — 2026

University of Gloucestershire · early completion with Merit

• Dissertation: a method for designing LLM prompts using conjoint analysis with AI agents — data-driven discovery and selection of prompt attributes.

Bahria University · 3.94/4.0 · Gold Medal, Summa Cum Laude

• Dissertation, “Smart Invigilator”, earned Pakistan government recognition and funding.

IEEE Open Access Journal

IEEE CDMA 2024, Saudi Arabia

IEEE Conference, Maldives

USASBE 2025, USA

Int. Journal of Computer and Systems Engineering

sole author, R

Guest speaker: AI & religious text in digital humanities.

On leadership and impact in AI consulting.

Interviewed on a leading, innovative approach as an AI consultant in the emerging field of Generative AI.

Acknowledged by EIN News and The Chicago Journal for contributions to ProAI's foundational Generative AI architecture.