Voice AI Engineering · 01
What Is AI and How Did We Get Here?
A builder's tour through AI history—from expert systems and machine learning to transformers—and why each layer matters for building Big Mama in production.
9 min read
Section
Building in Public
This is where I document the actual work — building voice and agentic AI products like Djembe AI, Big Mama, and Kelpy AI. Not the polished retrospective, but the process: what I tried, what broke, what I learned. Engineering deep dives included.
Voice AI Engineering
12 episodes-
-
Voice AI Engineering · 02
What Is an LLM? The Engine Behind Big Mama
A practical breakdown of how large language models work—tokens, embeddings, context windows, hallucination—and why the LLM is one engine inside a larger system, not the whole product.
10 min read -
Voice AI Engineering · 03
What Makes AI Agents Agentic?
The concrete difference between a chatbot that answers and an agent that acts—covering the agent loop, tool calling, memory, planning, and safety guardrails through a security engineer's lens.
9 min read -
Voice AI Engineering · 04
Why Voice-First AI Is the Future
Voice isn't just a feature—it's the front door: why conversation is the most natural interface and what it actually takes to engineer a voice-first AI that holds up in the real world.
11 min read -
Voice AI Engineering · 05
Building Big Mama with ElevenLabs: Voice Persona as Product Design
How to approach voice persona as a trust layer, not a cosmetic—covering TTS strategy, voice cloning ethics, latency tradeoffs, and what it means to design a culturally grounded voice from the ground up.
12 min read -
Voice AI Engineering · 06
Real-Time Voice Streaming with LiveKit: Building the Nervous System of Big Mama
Why voice AI demands live sessions instead of request-response cycles—and how LiveKit handles audio transport, latency budgeting, turn detection, and security for a production voice agent.
10 min read -
Voice AI Engineering · 07
Adding Memory and Planning to Voice AI: A Security Engineer's Perspective
Memory turns a stateless assistant into a relational agent—but only if you treat it as a privilege boundary with consent controls, tenant isolation, and a clear difference between memory and context.
9 min read -
Voice AI Engineering · 08
Connecting Calendars and Productivity Tools: Giving AI Agents Real-World Utility
When an AI agent connects to your calendar and task tools it stops being advice and starts taking action—here's how to design those integrations securely, with least-privilege permissions and per-action confirmation gates.
11 min read -
Voice AI Engineering · 09
Designing Human-Like Voice Experiences: Trust, Tone, and Transparency
Human-like voice AI isn't imitation—it's interaction that respects conversational rhythm, handles barge-in and repair gracefully, and stays transparent about what the system knows and does.
12 min read -
Voice AI Engineering · 10
Reliability, Latency, and Scaling Voice AI: Moving from Demo to Production
A demo works once; a product has to survive bad networks, slow APIs, user interruptions, and real business stakes—here's how to build a latency budget, design graceful degradation, and apply security engineering discipline to voice AI at scale.
9 min read -
Voice AI Engineering · 11
Building AI for Black Communities and SMBs
Black communities shouldn't only be consumers of AI—they should be builders, owners, and beneficiaries: the case for starting with real community friction, not the latest model, and why data stewardship is a non-negotiable design constraint.
11 min read -
Voice AI Engineering · 12
How to Become a Voice AI Engineer: A Builder's Roadmap
Voice AI engineering spans ML, real-time systems, security, and product design—here's the exact phased learning path, skill map, and public build strategy being used to make the transition from security engineering to frontier voice AI.
10 min read