Forward Deployed Engineer · Applied AI / ML Platform

I embed with your team, find the messy problem, and ship AI into production.

Software and ML engineer across applied AI/LLM, cloud platform, DevOps and delivery — RAG/LLM systems, production ML pipelines and sub-second APIs — plus a habit of working inside the customer’s world and coaching the team until it sticks.

Download résumé (PDF) Email me

Remote (US) · contract-ready · concurrent engagements

Selected work

Proof, not adjectives.

Representative outcomes from production AI/ML and platform work. Every figure below traces to the skills source of record.

Healthcare · Applied AI

Treatment.com — AI/ML product line

As CTO at Treatment.com, owned the AI/ML product line across its Merlin → Decision rebrand, serving clinicians, schools and partners.

Owned the AI/ML product line as CTO through the Merlin → Decision rebrand.
Built AI/ML serving clinicians, schools and partners.

Applied AI · LLM

RAG & LLM applications

Retrieval-augmented generation and LLM applications shipped into real use — from document corpora to production answer quality.

Built RAG pipelines with LangChain + Pinecone that lifted answer accuracy ~30%.
RAG over Confluence datasets; LLM RAG application work at Georgia Tech.
Fine-tuning with LoRA/QLoRA (GPT-style models) for finance and healthcare; evaluation harnesses with human feedback and BLEU/ROUGE.

ML · Recommendation

Production ML at scale

Models in production where the numbers are the point: engagement, revenue and risk, measured in real traffic.

Production recommenders at 10M+ daily predictions — +34% engagement and $18M incremental revenue.
Risk-classification models (~89% accuracy) across a $2B+ premium portfolio; $22M+/yr saved.
Sub-200ms ML-backed API microservices for credit decisioning and fraud detection.

MLOps · Inference

Model serving & MLOps

Getting models out of notebooks and keeping them healthy: serving, versioning, canaries and drift monitoring.

Cut inference latency ~40% with vLLM and 4/8-bit quantization; multi-GPU training with DeepSpeed / Accelerate / PyTorch Distributed.
MLOps on GitHub + Terraform + AWS (ECS/Lambda) with CI/CD, A/B and canary deploys, drift/bias monitoring — 90% enterprise adoption.
HIPAA-compliant deployment; reduced diagnostic assessment time ~45%.

Platform · Cloud

Cloud platform & infrastructure-as-code

Platforms teams actually adopt — reusable infrastructure and paved roads instead of one-off environments.

Reusable IaC templates (Terraform, CloudFormation, AWS CDK) cut onboarding from weeks to hours across 15+ groups; provisioning cycles −60%.
AWS breadth across banking, insurance, telecom and healthcare — compute, data, streaming, containers and ML services.
Designed and deployed enterprise microservices and serverless pipelines (99.9% reliability at terabyte scale).

DevOps · SRE

Reliability & delivery enablement

Shipping faster without shipping worse — pipelines, observability and the coaching that makes it stick.

CI/CD build/deploy time −40–50% and manual deployments −65%; sustained 98.2–98.5%+ uptime with HA clusters.
Facilitated postmortems with >40% reduction in incident recurrence; observability with CloudWatch, Grafana and ELK.
Trained and coached 1,000–1,500+ engineers; TDD raised deployment quality ~70%; enabled 50–90 teams at scale.

Skills

The toolkit.

Grouped by theme — selected and re-ordered per engagement, the same way each tailored résumé is built.

Applied AI / LLM: RAG (LangChain, Pinecone, Weaviate, Chroma) · fine-tuning (LoRA/QLoRA) · GPT / LLaMA / Mistral / Falcon · Hugging Face Transformers · prompt engineering · eval harnesses · vector databases · NLP
ML / MLOps: TensorFlow · PyTorch · Scikit-learn · DeepSpeed · vLLM · quantization (4/8-bit) · MLflow · Ray · A/B + canary deploys · drift & bias monitoring
Cloud & platform: AWS (SageMaker, Bedrock, EKS, ECS, Lambda, Glue, EMR, Redshift, Kinesis, DynamoDB, CDK, CloudFormation) · GCP Vertex AI · Azure ML · Docker · Kubernetes / OpenShift · Terraform
DevOps / CI-CD: Jenkins / CloudBees · GitHub Actions · GitLab CI · Bamboo · Azure DevOps · Artifactory · Puppet · SonarQube · TDD / BDD / ATDD (Gherkin, Cucumber, JUnit, Mockito, Playwright)
Data: ETL architecture · Spark · Kafka (AWS MSK) · Apache Flink · PostgreSQL · Redis · ElasticSearch · real-time streaming · data quality
Languages: Python · Java · Go · TypeScript / JavaScript · C++ · Scala · SQL
Product / delivery: product ownership & discovery · roadmaps / OKRs · agile coaching (Scrum, Kanban, SAFe) · stakeholder & executive communication · enablement at scale · Atlassian suite (Jira, Jira Align, Confluence)

How I work

Discover → build → hand off.

A forward-deployed loop: get close to the problem, deliver working software fast, then make the team self-sufficient.

Discover

Sit with the real problem and the people living with it. Requirements-gathering, product discovery and stakeholder alignment turn an ambiguous ask into something buildable.
Build

Ship the smallest thing that works, in the customer’s environment — AI/ML services, platform automation or delivery pipelines — then harden it for reliability and scale.
Hand off

Leave the team better than you found it. Documentation, runbooks and coaching so the solution is owned by the people who depend on it.

Credentials

Certifications.

AWS Certified Solutions Architect — Associate & Professional
AWS Certified Cloud Practitioner
AWS Certified Machine Learning — Specialty
Atlassian Certified Jira Administrator (ACP-100)
ICP-ACC Certified Agile Coach

Education

Degrees.

M.S. Decision Science — Georgia State University
B.S. Computer Science — Georgia State University
B.A. English (Logic, Rhetoric & Composition)

Languages: English, French, Arabic.

Contact

Let’s talk about the problem you’re stuck on.

Fully remote, contract-first (W2 / C2C / 1099). If a role demands exclusive full-time commitment, say so up front.

jabrealmoe@gmail.com Résumé (PDF)

GitHub X / Twitter