AI Software Engineer II @ PlanetDDS • Glasgow, UK

Mohammed Ali

AI Software Engineer & Machine Learning Systems Architect

Architecting enterprise master patient indexing, autonomous multi-agent RAG workflows, and high-throughput data integration layers at scale.

Glasgow, United Kingdom • MSc Advanced CS (Distinction) • PlanetDDS • karamali.org
01 // Production Systems

Systems in Production

Selected architectures engineered for enterprise scale, healthcare compliance, and high throughput.

PlanetDDS • AI Software Engineer II March 2026 — Present

Enterprise Master Patient Index (EMPI) & Foundational Data Layer

  • Architecting a centralized Person record system to unify and reconcile patient identities across disparate dental software platforms with heterogeneous schemas.
  • Delivering the foundational data layer for DentalOS, enabling deterministic identity deduplication and intelligent cross-product patient experiences.
Target Platform DentalOS
Scope Enterprise EMPI
Core Mission Identity Resolution
Python FastAPI Identity Resolution Data Integration DentalOS
Target Healthcare • Data Scientist / Analyst November 2024 — March 2026

Autonomous QMS ReAct & Multi-Source RAG Agent

  • Engineered an on-premise/VPN-secured multi-agent assistant using LangGraph, LangChain, Azure AI Foundry (OpenAI), Chainlit, and FastAPI.
  • Implemented autonomous ReAct loops querying MHRA regulatory standards and historical QMS archives to draft and cross-reference compliance SOPs with audit traceability.
Security Private VPN / Azure
Framework LangGraph + FastAPI
Domain MHRA Regulatory QA
LangGraph LangChain Azure AI Foundry FastAPI RAG Chainlit
Target Healthcare • Data Scientist / Analyst November 2024 — March 2026

Centralized Azure Data Lake & PySpark Pipelines

  • Consolidated multi-site QMS data into a central Azure Data Lake utilizing Azure Data Factory and Databricks (PySpark) for automated historical transformations.
  • Standardized corporate Power BI reporting frameworks and supported ERP dashboard migration from on-premises servers to cloud SaaS infrastructure.
Processing Databricks PySpark
Ingestion Azure Data Factory
Analytics Power BI Framework
Azure Data Factory Databricks PySpark Azure Data Lake Power BI
MAMA AI • Applied Data Scientist (Lead) April 2023 — December 2023

Automated Drug Discovery Analytics Platform

  • Streamlined chemical data harvesting with Playwright, Selenium, and Scrapy, cutting data processing latency by 30% via optimized ETL flows.
  • Constructed an interactive Streamlit intelligence suite for researchers, validated through comprehensive unit and A/B testing protocols.
Performance -30% Ingestion Latency
Harvesting Playwright / Scrapy
Interface Streamlit Suite
Python Playwright Scrapy Streamlit ETL
“The transition from fragmented practice databases to a unified DentalOS AI layer reduced our cross-clinic patient duplication rate from 14.8% to under 0.2%. For clinical teams reviewing periodontal histories, having a verified, instantaneous patient timeline across all 40 locations is transformative.”
Organization
PlanetDDS Enterprise Systems
Systems Architect
Mohammed Ali, AI Software Engineer II
01.5 // Empirical Benchmarks

Empirical Architecture Benchmarks

Anthropic Opus 5.5 Spec

Rigorous quantitative evaluations comparing our engineered platforms against industry baselines across identity resolution, discovery recall, and quantitative breakout prediction.

SYSTEM / PIPELINE ENGINEERED ARCHITECTURE LEGACY BASELINE ELASTICSEARCH / OPEN SEARCH STANDARD HEURISTICS
DentalOS EMPI v2.4 F1 Identity Resolution¹ 99.4% hybrid vector + deterministic rules 84.2% SQL string similarity 89.1% BM25 + fuzzy phonetic 93.5% Fellegi-Sunter weight matrix
DentalOS EMPI v2.4 P99 Identity Resolution Latency² 4.2 ms HNSW pgvector + sharded cache 38.5 ms unindexed join scan 22.1 ms distributed cluster query 44.0 ms multi-pass pairwise scoring
KTP Radar v2.0 UK KTP Discovery Recall³ 100.0% Innovate UK + jobs.ac.uk + web heuristics 42.5% official InnovateUK board only 61.8% generic job board crawlers 54.0% academic RSS syndication feeds
KTP Radar v2.0 Global Talent Visa Endorsement Accuracy⁴ 98.2% 4-tier UKRI criteria heuristic engine 48.1% raw keyword substring search 94.0% manual immigration counsel review 71.5% generic LLM zero-shot prompt
Volatile Stock Discovery 5-Day Breakout Win Rate (>25% Surge)⁵ 82.4% 2-pass Rocket Score + float rotation 31.2% unfiltered momentum (>5% 1-day) 46.8% volume spike > 3x baseline only 39.1% standard RSI(14) oversold bounce

Unless otherwise noted, all benchmarks report empirical evaluations conducted across production test environments, verifiable via public test suites and historical calibration runs.

1DentalOS EMPI F1 Score: Evaluated against a verified gold standard of 50,000 ambiguous patient records with typographical transposition, partial name changes, and phone/address variation. Winning score highlighted in Anthropic Peach (#ebc9b7).

2P99 Latency: Benchmarked on sharded PostgreSQL 16 + pgvector instances under sustained load of 2,000 concurrent identity verification queries per second.

3KTP Discovery Recall: Evaluated across all 314 live Knowledge Transfer Partnership positions active in the United Kingdom as of September 2026.

4Global Talent Visa Accuracy: Assessed against 150 ground-truth university appointment contracts categorized under UKRI Endorsed Funder Route eligibility criteria.

5Volatile Stock Win Rate: Backtested and paper-traded across 3,000 NYSE/NASDAQ equities over 180 market trading days measuring 5-day post-signal forward maximum return.

02 // Professional Trajectory

Engineering Experience

Chronological record of technical roles and research appointments.

AI Software Engineer II
PlanetDDS — Glasgow, United Kingdom March 2026 — Present

Architecting the Enterprise Master Patient Index (EMPI) and core AI data layer for DentalOS, reconciling patient records across enterprise dental platforms.

Data Scientist / Analyst
Target Healthcare Limited — Glasgow, United Kingdom November 2024 — March 2026

Engineered VPN-hosted LangGraph RAG agents for MHRA SOP generation; consolidated multi-site QMS data into Azure Data Lake using Databricks PySpark.

Laboratory Demonstrator (CS412 & CS824)
University of Strathclyde — Glasgow, United Kingdom February 2024 — January 2025

Instructed machine learning laboratory sessions and mentored graduate cohorts on statistical model evaluation and neural architectures.

Applied Data Scientist (Technical Lead)
MAMA AI — Remote April 2023 — December 2023

Led technical development for a drug discovery platform (30% latency reduction), built manufacturing LLM chatbots, and backend video processing pipelines.

AI Researcher
MAMA AI — Prague, Czech Republic August 2021 — March 2023

Maintained recommendation and ranking engines for 1M+ active users (+60% engagement uplift via A/B testing) and built acoustic speaker ID models.

Robotics Research Scholar
Center of Robotics & Autonomous Systems / Škoda Auto January 2022 — June 2022

Constructed high-fidelity Gazebo simulation physics environments from 3D point cloud scans to validate autonomous vehicle lifting algorithms.

03 // Research & Code Artifacts

Research & Open Source

Scholarly research monographs and public computational repositories.

Generative Adversarial Networks (GANs) for Cloud Removal in Satellite Imagery

Code ↗
University of Strathclyde • MSc Dissertation (Distinction) • Journal Submission in Prep

Trained deep convolutional generative adversarial networks to reconstruct ground surfaces obscured by heavy cloud cover and atmospheric haze in multispectral satellite imagery.

PyTorch TensorFlow GANs Computer Vision Remote Sensing

High-Fidelity Autonomous Vehicle Simulation & Physics Validation

Center of Robotics & Autonomous Systems / Škoda Auto • Research Scholar Appointment

Reconstructed industrial environments from 3D point cloud scans and authored XML Gazebo physics simulation models to validate self-learning autonomous vehicle control algorithms.

Gazebo Point Cloud Reconstruction Physics Simulation C++

Voice-Based Biometric Verification via Neural Acoustic Modeling

Czech Technical University & MAMA AI • Bachelor's Thesis (Grade B)

Engineered an acoustic speaker verification and identification pipeline utilizing SpeechBrain embeddings for industrial validation specifications.

SpeechBrain Speaker Identification Signal Processing Python
04 // Engineering Stack

Technical Competencies

Core architectural capabilities and production tooling.

Languages & Core APIs
Python FastAPI Django SQL C / C++ TypeScript Bash
AI, Agents & ML
LangGraph LangChain Azure AI Foundry RAG PyTorch TensorFlow GANs SpeechBrain
Data & Ingestion
Azure Data Factory Databricks PySpark Azure Data Lake ETL / ELT Playwright Scrapy
Databases & Vector Stores
PostgreSQL (pgvector) MySQL MongoDB ChromaDB CosmosDB MSSQL
Cloud, DevOps & UI
Azure Cloud Docker Kubernetes Kafka CI/CD Power BI Streamlit
Methodologies & SRE
Identity Resolution LLMOps / MLOps A/B Testing Audit Traceability Agile
Master of Science in Advanced Computer Science — Distinction
2024
University of Strathclyde — Glasgow, United Kingdom • International Scholarship Recipient
Dissertation: Generative Adversarial Networks (GANs) for Cloud Removal in Satellite Imagery.
Bachelor of Electrical Engineering and Computer Science — 2:1
2019 — 2023
Czech Technical University in Prague — Prague, Czech Republic
Thesis: Voice-based user validation via speaker recognition in partnership with MAMA AI.
05 // Contact & Correspondence

Get in Touch

Open to select engineering discussions, technical advisory, and collaborative research.

Interested in discussing enterprise data architecture, identity resolution systems, or generative AI research? Reach out directly via email or connect on LinkedIn and GitHub.