Sachin S.

Most AI consultants hand you a demo. I build the integration layer that connects Claude and Agentforce to the CRM, files and tools your business already runs on, so it still works in month six. 11+ years in cloud and platform architecture. Top Rated, 100% Job Success, 40+ Upwork jobs delivered. 𝗥𝗲𝗰𝗲𝗻𝘁 𝘄𝗼𝗿𝗸: - AI operations layer for a US commercial real estate brokerage: 8 hours of daily manual work down to under 20 minutes - Custom enterprise MCP server with its own UI, LLM agnostic, running against both Claude and ChatGPT custom connectors - Claude AI workflows wired directly into a live Salesforce org (Rethink+) - Agentforce Sales and Service agents with Slack, Outlook, Dropbox and Smartsheet connected through MCP 𝗪𝗵𝘆 𝗶𝘁 𝗵𝗼𝗹𝗱𝘀 𝘂𝗽 𝗶𝗻 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻: before AI, I built production cloud. A secure GCP landing zone in Terraform for an Indian bank covering IAM, network segmentation, WAF, regional failover DR and RBI compliance. An AWS to GCP migration of 100+ services with zero downtime at cutover. So my AI builds ship with access control, logging and a rollback plan, not an API key sitting in a config file. 𝗪𝗵𝗲𝗿𝗲 𝗜 𝗵𝗲𝗹𝗽 𝗺𝗼𝘀𝘁: - Custom MCP servers, enterprise grade, UI ready, LLM agnostic - AI agents that touch real systems: CRM, email, files, project tools, internal APIs - Agentforce Sales Agent and Service Agent builds - Workflow automation where AI handles the judgment step and the system handles the rest - Guardrails, access control and observability for AI in production - Architecture reviews before you commit budget to a build 𝗦𝗮𝗹𝗲𝘀𝗳𝗼𝗿𝗰𝗲 𝗶𝘀 𝘀𝘁𝗶𝗹𝗹 𝗰𝗼𝗿𝗲: most of my Upwork reviews come from it. Migrations (legacy CRM to Salesforce, org to org, Classic to Lightning, HubSpot to Salesforce) with a real rollback plan and zero data loss. Cleanup for over-customised, slow orgs. Advanced Flow, Apex, APIs and Data Cloud. 𝗛𝗼𝘄 𝗜 𝘄𝗼𝗿𝗸: 7 years leading delivery teams and pre-sales scoping. I size the work on the first call, flag the risks you have not priced in, and tell you when what you asked for is not what you need. Architecture before code. Security and documentation are part of the build from day one. Handed over so your team owns it. 𝗪𝗮𝘆𝘀 𝗰𝗹𝗶𝗲𝗻𝘁𝘀 𝘀𝘁𝗮𝗿𝘁, 𝗺𝗼𝘀𝘁𝗹𝘆 𝗳𝗶𝘅𝗲𝗱 𝗽𝗿𝗶𝗰𝗲: - MCP feasibility review and architecture - Custom MCP server build, one system - Agentforce Sales or Service Agent build - AI workflow automation sprint, two weeks - Salesforce migration assessment or org health check 𝗖𝗲𝗿𝘁𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 (𝟮𝟬+): Anthropic (9): Claude Code, MCP Advanced Topics, Agent Skills, Claude on Bedrock, Claude on Google Cloud and more AWS (5): Solutions Architect Professional, Security Specialty, SA, Developer, SysOps Associate Google Cloud (3): Professional Cloud Architect, Associate Cloud Engineer, Digital Leader Kubernetes (3): CKA, CKAD, CKS Salesforce: Platform Developer I, Agentforce Specialist Also: Terraform Associate, Copado (4), PMEC 𝗖𝗼𝗿𝗲 𝘀𝘁𝗮𝗰𝗸: MCP, Claude, Agentforce, GCP, AWS, Kubernetes, Terraform, Salesforce (Sales Cloud, Service Cloud, Data Cloud, Flow, Apex, APIs), Slack, Outlook, Dropbox, Smartsheet, ClickUp, Copado. Send me the system you want AI inside and what is failing today. I will tell you what it would actually take, and whether I am the right person for it.