I build the system,not the pitch.

I turn uncertain, operational problems into working AI systems, automations, and complete software products—directly for focused work, or leading a senior team when the scope is larger.

Base: Barcelona, Spain
Current: Founder, NovaNest Pro
Mode: Direct build / senior product team

Fig. 01 — Build pathControl stays visible

50+AI and automation systems / product builds

100M+Tokens used in applied model work during 2026

6+Years across engineering, product, and operations

1→teamOne accountable lead, senior depth as required

Selected systems.
Built to be used.

The common thread is not an industry or a model. It is taking responsibility for the whole path—from the messy operating problem to the software people can actually run.

2025—now NovaNest Pro
Founder / product / build An AI operator for home-service professionals: lead intake, business rules, reply control, follow-up, booking, and operational state.

2021—24 DataSpot
Product lead Product direction and delivery across a data-heavy platform—aligning technical decisions with what the business and its users needed next.

2021 Interview Simulation
Co-founder / builder An early applied-AI product for practicing the pressure, structure, and feedback loop of a real interview.

Ongoing 50+ product builds
R&D / client systems Agents, internal tools, automations, integrations, prototypes, and production systems across focused experiments and client work.

Featured case / NovaNest Pro

Built after losing leads while installing cameras myself.

Observed early outcome

First autonomous booking: $190.

Founder-account trial
No message approval at booking time

The system was shaped inside the work: incoming Thumbtack leads, incomplete details, pricing rules, availability, customer replies, and the risk of an AI taking the wrong action. The product keeps business state and human control visible instead of hiding them inside a chat.

Read the case →

6Real leads processed in the documented early trial

2Bookings recorded in that trial

1External professional trial recorded separately

Any software problem.
The team scales to the work.

I do not sell a narrow automation recipe. I can enter at the idea, the broken workflow, the difficult integration, or the half-built product—and own the route to a usable result.

New product

Problem framing, product decisions, architecture, interface, implementation, deployment, and the first real operating loop.

AI + automation

Agent workflows, model selection, retrieval, structured outputs, evaluation, approvals, retries, audit trails, and human handoff.

Integration + data

APIs, operational data models, event flows, CRM logic, third-party services, reporting, and the unglamorous edges where systems usually fail.

Rescue + hardening

Clarify an inherited codebase, remove fragile behavior, repair the critical path, add observability, and make ownership possible again.

One accountable lead.
More senior hands when earned.

For a focused build, I work directly. For a complete product or larger platform, I lead a small senior team across backend, infrastructure, AI reasoning, frontend, and product interface.

Direct build

Hovhannes

A focused prototype, automation, integration, rescue, or product slice—with the person making the decisions also doing the work.

Lean product team

Hovhannes + senior specialists

A complete product that needs parallel engineering depth without adding layers of coordination or junior handoffs.

Dedicated pod

Expanded around the system

For larger scopes, the team grows deliberately around architecture, frontend, infrastructure, QA, and the operating requirements.

Operator / product / engineering

Hovhannes Hovhannisyan

Problem definition, product judgment, system design, build leadership, and the final operating result.

Backend / infrastructure / AI

Anri Hayrapetyan ↗

Backend architecture, infrastructure, integrations, AI reasoning, and reliability.

Frontend / interface / product

Hakob Beglaryan ↗

Frontend systems, product interface, visualization, demos, and user experience.

Speed is a method.
Not a promise.

I move quickly because I reduce uncertainty early, use the strongest tools available, and keep the business decision close to the code. Production time still depends on access, data, risk, and scope.

First usable build A well-scoped product slice can often be made tangible in 48 hours. It is the start of evidence, not a production guarantee. Fast feedback

Model work 100M+ tokens used in 2026 across practical evaluation and building. New models enter the system only when they improve the measured task. Applied

Operational reliability Failures, retries, logs, approvals, edge cases, and ownership are designed into the working loop instead of treated as post-launch polish. Visible

Enterprise context Corporate client work includes companies whose own customer or partner portfolios reference NASA, Tesla, and Meta. Those organizations were not direct clients or endorsers. Indirect

If the problem is difficult,
it is probably worth discussing.

Send the operating constraint, the failed attempt, or the product brief. I will tell you what I think can actually be built.

hovhannes@novanestpro.com