Production Agents 2026
10 min read6 primary sources
MCP connects agents to tools. A2A coordinates independent agents, including through stateful tasks. Learn where each belongs and what neither protocol solves.
NewMCP vs A2A in 2026: A Production Architecture Guide
AI systems engineerToronto, Canada
ReliableAIstartsafter the demo.
I design retrieval, agent, and evaluation systems to stay grounded in evidence, bounded by clear controls, and measured against reality when the happy path ends.
Latest writingReliability loop
Evidence before confidence
01Ground
02Constrain
03Evaluate
AI for learningRetrieval systemsEvaluation
Latest writing
Ideas with an audit trail.
Field notes and long-form essays on building AI systems that are grounded, bounded, and measurable.
Production Agents 2026
13 min read8 primary sources
Build agent memory as governed state: control writes, preserve provenance, supersede stale beliefs, enforce deletion, and verify actions against evidence.
Production Agents 2026
12 min read5 primary sources
Treat prompt injection as a source-to-sink security problem. Contain the agent, limit authority, and gate consequential actions outside the model.
Production Agents 2026
11 min read5 primary sources
Evaluate the whole agent system: verify final state, inspect traces, control the environment, repeat trials, and turn real failures into regression tests.
Production Agents 2026
11 min read4 primary sources
Task horizon, uninterrupted runtime, and parallel agent-hours are different clocks. Learn how to read agent autonomy research without overstating it.
Research series · 5 parts
Production Agents 2026
Source-led guides to protocols, autonomy, evaluation, security, and memory. Read in order, or jump to the part you need.
Start with part 1 28 primary sources cited
PhD
Computer Science
MSc
Artificial Intelligence
RAG · Agents · Eval
Core focus
Toronto, Canada
Based in
What I work on
Systems designed to survive contact with reality.
The work spans the full reliability stack: finding the right evidence, coordinating model behavior, measuring quality, and translating research into products people can trust.
01
Retrieval & RAG systems
Designing retrieval-augmented systems that ground model output in real sources: hybrid retrieval (lexical + dense), semantic chunking, reranking, and citation-grounded answers.
02
Multi-agent orchestration
Building agent systems that plan, call tools, and manage state reliably, with deterministic guards, retries, and explicit boundaries instead of hopeful prompting.
03
Evaluation & reliability
Treating evaluation as engineering: golden datasets, faithfulness and groundedness metrics, regression gates, and adversarial testing so quality is measured, not assumed.
04
Applied ML & research
Bringing PhD-level research rigor to production: information retrieval, network analysis, and adversarial/applied machine learning, translated into systems that ship.
Start a conversation
Building something that needs to work beyond the happy path?
I'm always interested in thoughtful conversations about retrieval, agents, evaluation, and AI for learning.