Gaurav Kheterpal Portfolio
Salesforce MVP Hall of Fame member (MVP since 2016) whose blog tracks new platform architecture as it lands, including 2025 posts on Agentforce, A2A, and MCP.
Location
Somewhere on planet Earth, most likely near a laptop.
Skills
What makes it work
A breakdown of the choices that make this portfolio stand out.
Writing about emerging architecture as it becomes relevant
Covering Agentforce + A2A + MCP shows the blog tracks the current edge of the platform, not just legacy topics.
What Salesforce Developers can take from this
Specific, actionable tips to apply to your own portfolio — no generic advice.
Write early about genuinely new platform capabilities, not just tutorials on established features, to signal you're tracking the frontier.
Case studies
Hands-on AI-agent experiments and plain-spoken explainers on Salesforce credentials. Gaurav Kheterpal's blog mixes working code with architecture write-ups.
Multi-Agent Madness: Agentforce + A2A + MCP
An architecture explainer on how Agentforce acts as Salesforce's orchestration layer: A2A for agent-to-agent calls to Google, LangGraph, ServiceNow and SAP agents, and MCP for reaching enterprise systems through MuleSoft, Heroku and third-party servers. It closes with the practical risks: debugging distributed agents, multi-vendor security, and operational overhead.
It maps new, fast-moving protocols onto a platform clients already run and names the risks as well as the benefits, which is what an architect is paid to do.
Separates A2A (agent-to-agent) from MCP (agent-to-system) roles
Covers MuleSoft and Heroku as MCP hosting paths
Explicit section on debugging, security and vendor-overhead risks
Demystifying Partner Learning Camp & Accredited Professional Exams
A three-post series on Salesforce's Partner Learning Camp and its then-new Accredited Professional credentials: Part 1 (July 2021), Part 2 (September 2021) and a one-year review (August 2022). The review is organized as 'The Good', 'What needs to be fixed?' and 'What's really needs to be fixed ASAP!', covering content, vouchers, Examity proctoring, Trailhead sync and platform performance.
It gives candid, first-hand feedback on a vendor's credentialing program, backed by his own exam record, which shows he can evaluate a platform as well as use it.
Tracks a new credential program from launch to its first anniversary
Structured good / fix / fix-now critique
Backed by his own record: 21 of 34 AP exams passed
Projects
ADK Salesforce Connector (Agentforce + Google ADK experiment)
A Python experiment that connects a Google ADK agent to Salesforce so users can query CRM data in natural language, e.g. 'Show me all the open opportunities in New York.' It uses a Connected App for OAuth 2.0, a get_salesforce_access_token function, and a query_salesforce function that runs SOQL over the REST API. The ADK agent routes user input to these functions.
It is a working cross-vendor agent integration rather than commentary, and it shows he can wire LLM agent frameworks to enterprise data through standard APIs.
Natural-language questions translated into SOQL against a live org
Connected App + OAuth 2.0 setup documented step by step
Runnable Colab notebook linked from the post
More Salesforce Developer portfolios
Ready to build your portfolio?
Follow the Salesforce Developer roadmap — skills, projects, and timeline to get hired.