AI engineering
Hire senior AI engineers who ship agents to production.
Our AI and Python engineers join your Slack, Jira and Git within 48 hours. They build agents, RAG pipelines and copilots wired into your databases and APIs, then a senior engineer reviews and tests every one before it goes live.
- Kickoff
- 48 hours
- Full-time from
- $3,400/mo
- Senior from
- $6,600/mo
What we deliver
What our AI engineering team delivers
Production AI agents
Agents that complete multi-step work such as invoice processing, report generation and escalation routing, with tool permissions scoped per role.
RAG on your documents
Retrieval over contracts, manuals and tickets using pgvector or your existing database, with source citations in every answer.
MCP servers and tool layers
Model Context Protocol servers that expose your systems to Claude, ChatGPT and in-house agents with OAuth 2.1 and per-tenant limits.
AI sidecar for legacy apps
An assistant beside an existing SQL and web UI application, so users get answers without a rewrite of the core system.
Evaluation and red-teaming
Test sets, prompt-injection checks and cost ceilings that every agent passes before deployment.
Observability and cost control
Tracing, token spend and policy alerts in production using AgentWire and OpenTelemetry.
Technology
- Python
- LangGraph
- LangChain
- FastAPI
- FastMCP
- OpenAI
- Claude
- pgvector
- OpenTelemetry
- .NET
- Semantic Kernel
How we ship AI
Five gates before any agent reaches production.
- 1
Assess
Map data, risks and the workflow worth automating. One to two weeks.
- 2
Build
Agent, tools and retrieval wired to your SQL Server, Oracle or Postgres and APIs.
- 3
Senior review
An engineer with production experience reads every prompt, tool and permission.
- 4
Pre-deploy test
Evaluation sets, red-team prompts and cost ceilings before go-live.
- 5
Observe
Tracing, token cost and policy alerts in production, with a full audit trail.
Open source from our engineering lead
The same engineers who staff your pod build these in public.
github.com/qmmughal →- UpgradePilotAI agent pipeline that upgrades ASP.NET Zero projects to a target version..NET
- AgentWireObservability, security and cost analytics for AI agents, on ASP.NET Core..NET
- dotnet-agent-frameworkLightweight multi-agent orchestration for .NET..NET
- production-mcp-serverFastMCP, FastAPI and LangGraph with OAuth 2.1, pgvector and OpenTelemetry.Python
- SQL RecycleBinCaptures deleted and updated SQL Server rows so you can query them back.SQL
- smart-slm-routerOpenAI-compatible proxy that sends simple prompts to small models and escalates the rest.Python
Related work
Proof from a real engagement
See how we delivered similar work, or browse all case studies.
BlueberryCMS · United States
a 7-year dedicated team that re-architected a legacy CMS.
BlueberryCMS has worked with a dedicated 92 Logics team for seven years. The team replaced a monolithic CMS with modular .NET Core services and a React front end, and continues to build and support the platform.
Read case study →FAQ
Common questions
What does a 92 Logics AI engineer cost?
A full-time mid-level AI engineer starts at $3,400 per month and a senior AI engineer or tech lead at $6,600 per month. Fixed-scope agent builds start at $8,000 for a proof of concept.
Can your AI engineers work with our existing .NET system?
Yes. Most of our AI work sits beside .NET and SQL Server systems, often as an AI sidecar that reads the existing database without changing the core application.
How do you stop an agent from doing something wrong in production?
Each agent gets least-privilege tool access, human approval gates for irreversible actions, evaluation tests before release, and tracing with alerts after release.
Which models do you work with?
OpenAI, Anthropic Claude, Azure OpenAI, and open models served on your own infrastructure when data must stay on-premises.
Related services