Go (Golang) Backend Developer with 3 years of experience in designing, developing and optimizing highly loaded, distributed systems and microservice architecture. I have deep knowledge of the entire development stack, from API design to working with databases, message brokers, and containerization. Key expertise: Golang: Using frameworks (Fiber, gRPC) to create high-performance, parallel applications. Databases and repositories: PostgreSQL, Redis, ClickHouse, optimization of complex SQL queries and migrations. Architecture: RESTful API, gRPC, Microservices, WebSocket, Event-Driven Architecture (Kafka). Infrastructure: Docker, Docker Compose, Kubernetes, CI/CD. Tools: Swagger, Jaeger, Prometheus, Casdoor/IAM (OAuth 2.0).

I work with startups to build and maintain the infrastructure behind their products. Over the last 5+ years I've been responsible for production environments running on AWS and Azure, Kubernetes platforms, Terraform infrastructure, CI/CD pipelines, and the day-to-day reliability of customer-facing applications. My work ranges from reviewing existing infrastructure and improving deployment pipelines to troubleshooting production issues, automating cloud infrastructure, and helping engineering teams ship faster with confidence. Recently I've been working on a Kubernetes-based developer platform where I help build and operate multi-tenant infrastructure, improve platform reliability, develop Terraform modules, and solve distributed systems problems. I also manage AWS infrastructure for a healthcare SaaS platform, including EKS, GitHub Actions, Terraform, observability, and production operations. If you already have an application and need someone to improve the infrastructure behind it, I'd be happy to help.

I'm an experienced ex-Meta senior software engineer specializing in designing, deploying and integrating agentic AI in a robust, safe and reliable way. I also can do all the change management lifecycle and build complex systems across your organization. I'm able to build and manage project from start to finish, while communicating on the progress in the meantime.

Punctual, sociable, quick learning. I am fond of philosophy and cinema, I like to learn something new, make new acquaintances. I want to become a professional in my field, so I spend a lot of time at self-study.

AI Engineer & Data Automation Specialist I build production-ready AI agents and data platforms that solve real business problems. With 6 years of data engineering experience and 1 year focused on Agentic AI, I ship systems that run in production—not just demos. What I Deliver: AI Automation & LLM Integration: Custom agents for customer support, internal data analysis, and document retrieval (RAG). Data Engineering & BI: End-to-end pipelines (ClickHouse, Snowflake, BigQuery) and dashboards (Power BI, Tableau, Metabase). Scalable Systems: Solutions serving millions of users with 95%+ CSAT and rigorous evals/tracing. Recent Production Impact: Customer Support Agent: Handles 10,000+ daily conversations for a 4M+ user platform. Fintech Data Bot: Acts as the main analyst, writing and running SQL for regulated data. HR Platform: Built solo in 5 weeks using parallel AI coding sessions. Stack: Python, TypeScript, LLM APIs (Claude, OpenAI), PydanticAI, Vercel AI SDK, SQL, Airflow, dbt. Freelancing on Upwork since 2021 with 100% Job Success and 35+ clients. Tell me what you want automated, and I will tell you honestly if an agent is the right tool.

I cut a manufacturing plant's energy costs by $5,000-6,000/month by modeling its chiller, pump, cooling tower, and AHU flow, predicting room cooling demand, and built the agentic AI assistant its executives now use daily to get answers from company data instead of waiting on a report. I build agentic systems, AI that reasons through multi-step problems and pulls in the right data or documents on its own, instead of one-shot chatbots that guess. Backed by a background in physics-based modeling and operations research most AI agent engineers don't have. Recent work: → HVAC optimization system covering the full chiller, pump, cooling tower, and AHU flow: predicts cooling demand per room and adjusts chiller staging and pump frequency against variables nobody controls, like scheduled machine activity and outside air conditions, cutting energy costs by $5,000-6,000/month → Virtual Data Analyst Studio, a multi-agent platform on OpenAI's and Claude's models: a supervisor agent routes business questions to expert sub-agents, several of which run through MCP servers connecting to internal sales and inventory databases plus public sources like OpenFDA, for deeper analytical lookups than a single chat response could give → Multi-modal RAG agent that reads screenshots and flowcharts to answer step by step instead of dumping a PDF on the user → Real-time voice interview system (GPT Realtime) with two-way conversation: HR can add questions mid-interview live, a CV screening step sets each candidate's background and experience level, and the AI adjusts its questions for a fresh grad versus an experienced hire, only scoring once a set of HR-defined objectives has actually been covered. Handles ~600 candidates/month at roughly $200/month in running cost. → PII detection system for a national bank: custom named entity recognition and POS tagging hitting 98% F1-score, with inference optimized through ONNX GPU, TensorRT, and OpenVINO for a 13% latency cut in production What I can help with: • Agentic workflows, multi-agent orchestration, and MCP-based tool integration for data, support, and operations • LLM integration across OpenAI, Claude, Gemini, and Azure OpenAI • API integration and production deployment (FastAPI, Docker, cloud) Tools I work with: LangChain, LangGraph, OpenAI/Claude/Gemini APIs, Azure OpenAI, MCP, Milvus, PostgreSQL, Snowflake, ClickHouse, FastAPI, Docker, Langfuse, Git. I also work with computer vision (YOLO) and NLP systems (NER, data extraction) when a project calls for it, though agentic systems are where most of my recent work lives. A quick note on availability: I work full-time as a data scientist, so I take on Upwork projects evenings and weekends, Indonesia time (GMT+7). I'd rather be upfront about that pace than promise round-the-clock turnaround I can't deliver, but you can expect a reply within a few hours during my working windows.

📚 A wealth of experience as a Lead Full-Stack Software Engineer at your service! With a career spanning major corporate giants like Tesla, The Home Depot, and HEB Grocer, my background gives me deep insights into building scalable, efficient, and innovative software solutions for businesses of all sizes. ⭐ Here’s a snapshot of what I can bring to your project ⭐ ✅ Proven Expertise: Skilled in React/NextJS, Tailwind CSS, GoLang, with a focus on architecting clean, maintainable, and efficient codebases. That's just the current tech stack, but I've worked with other technologies including Java, Angular, NodeJS, GraphQL, Kubernetes, Pub/Subs, functions/agents, Terraform, CI/CD... ✅ Versatility: Experience with a wide range of projects, including e-commerce platforms, internal business tools, and last-mile delivery solutions. ✅ Problem Solving: Adept at tackling complex business challenges with well-thought-out solutions and scalable architectures. ✅ Team Leadership: I lead a team of vetted, highly skilled developers ready to help you achieve your goals with precision and efficiency. ✅ Attention to Detail: From understanding requirements to deployment, I ensure every line of code aligns with your vision and business objectives. ⭐ Why choose me over other freelancers? ⭐ ✅ Client-Centric Focus: I prioritize your needs, delivering solutions that truly add value to your business. And, provide you with valuable industry insight to help you deliver your product effectively. ✅ Clear Communication: I keep all lines of communication open, ensuring you're informed about every stage of your project. ✅ Empathy and Understanding: I approach every project with kindness and respect, ensuring your vision is met with the care it deserves. 🏆 What others have said about my work 🏆 ⭐⭐⭐⭐⭐ "Mike is a great engineer to work with. In his first week with the team, Mike was able to understand the project very quickly and start contributing to it. Mike is always open to listening to different ideas on how to solve problems which makes bouncing ideas with him effortlessly. He is great to work with and a great team member." – Chirag P. ⭐⭐⭐⭐⭐ "Our e-commerce platform needed a complete overhaul, and Mike delivered beyond expectations. The performance improvements alone were worth it." – Chris D. ⭐⭐⭐⭐⭐ "Mike is a go-getter. He helped us stand up the all the systems for our e-commerce business. That includes the Wordpress site, the Amazon account setup and automated delivery integration." – Luis M. 👨‍💻 Let’s Bring Your Vision to Life! Whether you need a small concept MVP brought to life, optimized internal tools, an off the shelf solution or all the way to a custom solution - my team and I are here to make it happen. We’re ready to tackle your toughest challenges and/or bring your next big idea to life. ➡️ Let’s connect and get started today!

I build production-ready AI + RevOps systems around Salesforce/HubSpot— orchestration, fine-tuning, MCP/semantic layers, RAG, and custom apps—so your sales & marketing teams ship faster and close more. 15 years in GTM systems; expert-vetted; 100% JSS; enterprise-grade delivery. What I Do AI Orchestration & Agents: Workflow design, tool use, evaluation, guardrails. Model Work: Prompt libraries, fine-tuning, preference optimization, evals. MCP & Semantic Layer: Domain schemas, tool servers, retrieval interfaces. RAG over GTM Data: Vectorizing Salesforce/HubSpot, docs, tickets, logs. Custom Apps & Integrations: Node/TS, React, FastAPI; SFDC/HubSpot apps; data pipelines. Signature Offers (fixed-scope starters) RAG QuickStart (2–3 weeks) – Ingest SFDC/HubSpot + docs → eval’d RAG with dashboards and red-team tests. MCP + Semantic Layer Blueprint (10 days) – Define tools, schemas, safety, and a reference MCP server. RevOps AI Tune-Up (2 weeks) – Lead/opp hygiene, scoring, routing, SDR copilot prompts + analytics. Salesforce/HubSpot AI Add-Ons – Case summarization, auto-notes, email drafting, knowledge search. Fine-Tune Sprint – Data curation, SFT or adapter-based finetune, eval harness, rollback plan. Recent Wins (abridged) Sales SDR Copilot: triage + drafting on HubSpot → 35% faster replies, 12% lift in meeting creation. SFDC Knowledge RAG: MCP tools + policy guardrails → deflection up 18%, CSAT +0.4. Quote Assist: approvals + terms via LLM tools → cycle time −27%, fewer reworks. GenAI Data Layer: semantic layer on CRM + docs → GTM analytics in minutes, not days. Tech & Stack LLMs/AI: OpenAI, Anthropic, HF; RAG Semantic/MCP: Model Context Protocol, tool servers, function/tool calling Data: Postgres, ClickHouse, Redis, dbt, Snowflake, SQL, etc. Apps/Infra: Node/TS, React, Docker, GCP/AWS/Azure, any and all APIs CRM: Salesforce (Admin, App Builder, Sales/Service Cloud), HubSpot (RevOps/Automation) CRM App Dev: Custom and marketplace apps for Salesforce and HubSpot Who I’m Best For B2B SaaS, GTM-heavy orgs, public sector/regulated, field ops—teams with Salesforce/HubSpot that want measurable, production AI (not demos). How We Work Consultation (60–90 min): Objectives → success metrics → data/readiness check. Design Sprint: Architecture, safety/guardrails, eval plan. Pilot: Ship something that moves a KPI; then scale. Credentials / Certifications: MBA (Operations). Salesforce Data Architect, Admin, App Builder, Sales/Service Cloud. HubSpot Certified. (Full list on request.) Message me with: your CRM (SFDC/HubSpot), the GTM KPI to move, and target timeline. I’ll suggest a 2-week pilot that proves value fast.

AI projects rarely fail because of the model. They fail because the system around the model is weak — poor retrieval, fragile integrations, unstructured data, and no clear path to production. I help businesses build production-ready AI products from end to end. I’m Muhammad Hassan, an AI Engineer specializing in RAG systems, LLM applications, AI agents, document intelligence, backend development, and workflow automation. I build complete solutions that move beyond prototypes: systems that retrieve the right information, generate reliable outputs, connect with existing tools, and fit into real business operations. What I can help you build RAG chatbots and knowledge assistants I design end-to-end retrieval systems with document ingestion, chunking, embeddings, hybrid search, reranking, memory, and source-grounded responses. LLM-powered applications I build AI features and full applications for support, research, reporting, internal operations, and domain-specific workflows. Document intelligence pipelines I turn PDFs, scanned files, tables, reports, and other unstructured documents into structured, searchable, usable data. AI agents and workflow automation I create agentic workflows and automations using APIs, webhooks, and tools like n8n to reduce manual work and improve operational efficiency. Backend and product development I build the infrastructure behind AI products: Python services, FastAPI backends, vector database integrations, SQL pipelines, and deployment-ready APIs. Tech stack Python, FastAPI, LangChain, OpenAI, Claude, Llama, Mistral, Pinecone, Weaviate, ChromaDB, Supabase, SQL, ClickHouse, OCR, document parsing, API integrations, and n8n. How I work I approach every project like a product, not a one-off task. That means: clear architecture reliable retrieval and outputs maintainable code scalable design practical business value If you need a developer who can build the AI layer, backend layer, and automation layer into one complete solution, I’d be glad to help.

I build data pipelines and the LLM features that run in a production environment. ETL and warehousing, Machine Learning, RAG, document/data extraction. Backend + ML + data engineer, fluent English, US-timezone-friendly. - Data pipelines & warehousing - RAG over your docs - Document & data extraction - LLM Integrations Stack: Python, FastAPI, PostgreSQL/pgvector, ClickHouse, Snowflake, Dagster, Qdrant, Redis, Docker, Claude & OpenAI APIs. I'm an engineer by training, so I build for what happens when things go wrong: validation, fallbacks, and a human in the loop wherever a mistake would be expensive. I work async, communicate clearly, and ship in days, not months. Send me what you're trying to automate and I'll tell you honestly whether AI is the right tool for it.