Capabilities / The full working stack

From the agent’s next move
to the system around it.

I work across agent behavior, knowledge, conversation, verification and infrastructure. Explore the concrete capabilities below and the projects that put them to work.

01 / Capability

Agent harnesses & tooling

Build the execution layer that turns a model into a useful worker: its instructions, tools, reusable skills and boundaries. The harness makes the work repeatable and its outcomes inspectable.

Agent harnesses
Execution scaffolding
Reusable agent skills
Tool use
Prompt engineering
LLM and API integration

In the workKnight CTO agent engineering

02 / Capability

Loops & autonomous workflows

Engineer the cycle of reasoning, action, observation and revision. Combine model judgment with explicit execution controls, and carry a multi-step task through interruptions and changing evidence.

Looping agents
Hybrid agentic loops
Autonomous systems
Multi-step workflows
Bounded execution
Outcome verification

In the workAutonomous delivery and customer workflows

03 / Capability

Swarms & multi-agent systems

Give agents distinct jobs and the infrastructure to work independently. Design delegation, routing and scaling around clear ownership, explicit contracts and isolated workspaces.

Agent swarms
Multi-agent coordination
Delegation
Agent routing
Boundaries between agents
Agent scaling
Cloud agents
Isolated workers

In the workKnight CTO autonomous acceptance swarm

04 / Capability

Self-healing software

Build software-delivery loops that discover defects, implement repairs and bring a separately tested candidate to a release decision. Keep the original requirements and the actual tested revision connected throughout.

Defect discovery
Autonomous code repair
Candidate rebuild and deployment
Independent acceptance retesting
Evidence-gated promotion
Repair-loop recovery

In the workBounded repair → rebuild → browser retest → promotion proof

05 / Capability

Evaluation & agent reliability

Make correctness observable. Evaluate model outputs and complete workflows, investigate failure paths, and require evidence for the behavior a system claims to deliver.

AI evaluations
Guardrails
Automated test suites
Browser acceptance testing
Playwright / browser tooling
Criterion-level evidence
Release gates
Operational failure handling

In the workSwarm acceptance work and retrieval evaluation tooling

06 / Capability

Retrieval & grounding

Connect answers to source material, then test the retrieval layer itself. Work across dense documents, vector and hybrid retrieval, and checks that distinguish relevant evidence from confident-looking output.

Retrieval-augmented generation (RAG)
Vector search
Hybrid retrieval
Dense-document retrieval
Confidence calibration
Abstention checks
Embedding anisotropy
Relevance and corpus coverage
Citation linkage

In the workMCP Retrieval Probe, Atlas and legal AI

07 / Capability

Knowledge & agent memory

Build the pipeline between changing source material and an agent’s usable knowledge. Retain source revisions and exact passages, refresh embeddings incrementally, and keep retrieval tied to the reader’s access.

Source ingestion and reconciliation
Document chunking
Local embedding models
Incremental indexing
Source and passage provenance
Knowledge coverage checks
Conversation capture
Persistent conversation memory
Permission-aware retrieval

In the workAtlas knowledge and conversation infrastructure

08 / Capability

MCP & connected tools

Build and integrate MCP tools, connect agents to external systems, and inspect what those connections actually return. Expose useful engineering checks as both a CLI and tools an agent can call.

MCP server development
MCP integration
Agent tool interfaces
CLI evaluation tooling
Server-specific probe recipes
Endpoint-count consistency
OAuth discovery checks
REST API integration

In the workPublic MCP Retrieval Probe project

09 / Capability

Voice agents & avatars

Created Atlas at Knight CTO: an AI avatar with a conversational and knowledge system behind it. The engineering spans meeting media, speech, turn state, grounded answers and the lifecycle of a live session.

Conversational avatars
Tavus avatar integration
Recall meeting integration
Speech-to-text pipelines
Text-to-speech delivery
Streaming answers
Turn-taking and interruption controls
Meeting context
Media lifecycle and teardown
Audible-response latency investigation

In the workAtlas avatar and meeting-system engineering

10 / Capability

Messaging & response agents

Built a Slack auto-responder that reasons over the conversation before deciding whether to reply. Connect event handling to fresh context, human handoff and a durable record of delivery.

Slack auto-responders
Socket Mode event ingestion
Thread-aware context
Agentic relevance judgment
Durable event queues
Replay deduplication
Human reply precedence
Shared send ledgers
Uncertain-send reconciliation
Assistant authorship and reply controls

In the workKnight CTO Slack responder engineering

11 / Capability

Durable orchestration & state

Give long-running AI work a lifecycle beyond a single model call. Separate workflow control from stored state and evidence, retaining enough identity and ownership to recover deliberately.

Temporal
PostgreSQL state
Durable workflows
Task ownership
Versioned task/result contracts
Interruption recovery
Retry and cancellation control
Retained execution evidence

In the workAutonomous swarm and agent infrastructure

12 / Capability

Inference & model infrastructure

Work on the model execution underneath the agent: routing, serving and local inference. Make provider selection explicit and investigate the latency, capacity and cost of the actual workflow.

LiteLLM
vLLM
Local inference
Model serving
Model routing
Multi-LLM systems
Provider integration
Inference cost and latency analysis

In the workKnight CTO inference work and SettleOn multi-LLM agents

13 / Capability

Cloud & execution infrastructure

Build the environment agents run in, with isolated execution, repeatable infrastructure and explicit deployment ownership. Bring enterprise distributed-systems experience to the operational work around AI.

Kubernetes / k3s
AWS
Google Cloud / GCP
Docker
Terraform / infrastructure as code
Linux operations
CI/CD and GitHub Actions
Release engineering
PostgreSQL and Redis
Distributed systems

In the workKnight CTO systems and AWS engineering experience

14 / Capability

Controlled customer deployments

Build around the customer’s data boundaries and the decisions people retain. Connect access controls and auditability to the workflow, including supervised AI in sensitive professional settings.

Customer-controlled model keys
Access controls
Audit trails
Zero-data-retention configuration
Human supervision
Audience and source boundaries
Explicit action authorization
Operational support

In the workCustomer AI deployments, Atlas and responder controls

15 / Capability

AI products & automation

Take an idea through customer discovery, architecture, implementation and real use. At SettleOn, led a 10-person team and brought AI products to paying users; internal GTM agents supported the business behind them.

Customer discovery
Product architecture
Product development
Form-filling automation
Internal sales and GTM agents
Customer integration
Deployment and onboarding
Product launches
Ongoing support

In the workSettleOn products and Knight CTO client delivery

16 / Capability

Hands-on engineering & leadership

Stay close to the code while leading the decisions and people that make delivery possible. Combine daily engineering in Python, Rust and SQL with founder experience and engineering leadership at AWS and Workday.

Python
Rust
SQL
Architecture and implementation
Engineering leadership
Technical hiring
Team development
Cross-functional programs
Founder-led delivery

In the workKnight CTO, SettleOn, AWS and Workday