Isaac Obo Enimil  Pronounce my name

Kumasi, Ghana

• Designed a scalable AWS environment using Terraform and Docker, automating the deployment of a production-ready infrastructure featuring a secure, multi-tier VPC and modular application layers.
• Engineered a cost-optimized CI/CD engine by configuring Jenkins with dynamic EC2 Spot Agent provisioning, auto-scaling compute resources based on build queue depth to accelerate developer feedback loops and improve continuous integration throughput.
• Deployed a comprehensive observability stack utilizing Prometheus, Thanos, Loki, Grafana, and Jaeger — implementing distributed tracing, long-term metric retention, and automated alerting to provide 360-degree visibility and minimize MTTR, while exposing self-service dashboards that give developers on-demand insight into their environments and error logs.
• Fortified platform integrity by integrating AWS GuardDuty and CloudTrail, enforcing automated threat detection and immutable audit trails to maintain a robust security posture and "least privilege" governance.
• Operated containerized workloads on Kubernetes in production, right-sizing pod resource requests and limits to eliminate scheduler insufficient-capacity failures caused by over-provisioned deployments, restoring reliable pod placement and reclaiming wasted node capacity.
• Implemented GitOps delivery with Argo CD, treating declarative manifests in Git as the single source of truth for cluster state, and injected runtime credentials from HashiCorp Vault to keep secrets out of images and repositories.

California, United States (Remote)

• Engineered high-fidelity SFT (Supervised Fine-Tuning) datasets to optimize model performance for agentic reasoning and MCP tool-calling environments.
• Exceeded client KPIs for data accuracy and volume by synthesizing complex instruction-tuning pairs emphasizing multi-step reasoning and precise workflow orchestration.
• Created robust correction datasets by auditing model failures and transforming them into optimal reasoning paths, improving model self-correction capabilities.
• Delivered Round 11 results that propelled the team to #1 in dataset quality and operational throughput, based on client-verified benchmarks.

Toronto, Ontario, Canada (Remote)

• Built an end-to-end RAG system using FastAPI and a Zilliz vector store to reduce LLM hallucinations by grounding responses in verified client documents.
• Optimized document retrieval by implementing a microservices architecture handling embedding generation, indexing, and metadata storage via Firebase.
• Implemented a serverless ingestion pipeline to automate the transformation of unstructured data into searchable vector formats.
• Integrated AI models into a React frontend, allowing interactive, context-aware document interrogation and data extraction.