Technical due diligence and AI advisory
Get AI right.
Technical due diligence for VCs.Applied AI architecture for companies.
- 00
Interface
Chat, UI, prompts
- 01← Wrappers stop here
Model calls
Calls to a foundation model
- 02
Orchestration
Agents, retrieval, evals
- 03
Integrations
Customer APIs and data
- 04← Hard to replicate
Systems of record
ERP, CRM, core ledgers
- 05
Regulation
Compliance, audit, guardrails
01
For investors
A clear technical answer before the term sheet.
We review the code and architecture and talk to the engineering team. The report covers technology and people, not financials.
Technical due diligence
A review of the architecture, code, infrastructure and engineering team before you invest. We separate what runs in production from what only works in a demo.
Deliverable: Written report, red flags first, and a call to walk through it
Depth and lab-risk assessment
Could OpenAI, Anthropic or Google ship this as a feature? We check how deep the product sits in the customer's operations: the systems of record it writes to and the compliance work it carries. A model provider would have to rebuild all of that to compete.
Deliverable: A depth assessment that names the features most at risk
Portfolio support
Architecture reviews for portfolio companies facing a large technical decision, such as a platform rebuild or a choice of model provider.
Deliverable: Review sessions and written recommendations
02
How a review works
Four steps from scope to report.
Step 01
Scope
We agree on the questions that matter for your decision and the access we need.
Step 02
Access
Repository, architecture documents, and conversations with the engineering team.
Step 03
Review
We review the system and test the claims in the pitch against what we find.
Step 04
Report
A written report in plain language, and a call to discuss it with the partners.
03
For companies
AI built into the systems you already run.
We help teams pick the use cases worth building and design them around their existing systems. If needed, we lead the engineering work.
Strategy and architecture
Which use cases to build first and whether the data supports them, followed by an architecture for scaling, compliance and deployment.
Engineering
RAG pipelines, agents with evals, MCP servers, and integrations with internal APIs, knowledge bases and approval steps.
Fractional CTO for AI
Architecture decisions, code review, hiring and vendor choices. Engagements of 3, 6 or 12 months.
04
Background
Fifteen years on the engineering side.
Proferent is run by Gautam Dhameja, an engineering leader who has spent 15 years designing complex systems and explaining them to the people who fund and buy them.
- Former Director of Solution Engineering, Parity Technologies
- Cloud and IoT architecture, Microsoft
- Co-author of a published technical book on blockchain
- Executive education at Oxford Saïd and MIT Sloan
05
Get in touch
Send us the deal, or the problem.
A short note is enough: the company, the stage, and what you need to know. We set up a 30-minute call. If it is a good fit, you get a clear scope and a fixed price within a few days.