Data & AI Mentor at Insignia · Corporate Trainer & Mentor in AI Programming, LLMs & RAG.
Former Head of Data Science at NoLimit Indonesia · 350+ sessions · 7,000+ learners · 20+ companies.
Hi! I'm Fahmi's AI assistant. Ask me about my training programs, experience, or how I can help your team learn Data & AI.
Trusted by 20+ companies & 10+ academic partners
Hands-on Data & AI training, customized to your industry and team's starting point.
I'm Fahmi, a Data Scientist and AI/NLP Engineer. Currently Data & AI Mentor at Insignia, driving talent transformation and technical mentorship in Data Science, AI, NLP, and LLMs across multiple business units. I help companies and individuals go from “we should use AI” to working automations, dashboards, and models they can use at work right away.
Before focusing on training, I spent 4+ years at NoLimit Indonesia as Head of Data Science, building and leading production NLP, LLM, and RAG systems for large-scale social media monitoring. That production experience shapes how I teach: every program is built around real use cases, hands-on from the first session, with clear outputs at the end.
Since 2019 I have delivered 350+ sessions for 7,000+ learners across 20+ companies and 10+ universities.
Built production NLP, LLM & RAG systems at scale.
350+ hands-on sessions for corporate teams and universities.
End-to-end programs, labs, and assessments.
MAY 2026 - PRESENT
Drive talent transformation, curriculum development, and technical mentorship across Data Science, Data Engineering, AI, NLP, and LLMs for multiple business units, designing and delivering training programs that build technical capability across the organization.
JAN 2019 - PRESENT
Delivered 350+ training sessions on AI Programming, AI, Machine Learning, NLP, and Data Science for professionals, students, and executives. Specialized in AI model building, deployment, and applied NLP systems using Python, PyTorch, TensorFlow, and FastAPI. Designed enterprise programs on Building Enterprise AI Chatbots with LLM & RAG (architecture, vector databases, prompt engineering, AI security & governance, Docker/Kubernetes deployment), plus developer-track programs on LangChain, RAG, ReAct agents, n8n workflow automation, multi-agent orchestration (CrewAI, AutoGen), and MLOps (MLflow) via PT Smartindo Manajemen Informatika. Clients include Bank Danamon, Toyota Astra Motor, Bayer, PLN, Freeport, Jasa Marga, Bank BSI, Seabank Indonesia, Pegadaian, and Ortax / PT Integral Data Prima, with academic engagements at UI, UGM, IPB, Atma Jaya, UPN Yogyakarta, ULBI, and Politeknik Madiun.
A curated selection of enterprise-grade NLP, predictive modeling, and AI automation solutions delivered for high-scale environments.
Responding to public opinion depends heavily on timing. Late information makes decisions less relevant, and a periodic-report approach means new issues are only spotted once the situation has already escalated.
Real-time visibility into public sentiment, more proactive decision-making, and earlier containment of escalation risk. The system stays stable even under sudden traffic spikes.
Taught in: Intelligent AI Agents for Workflow Automation · Data Analytics & Visualization
Daily operational reports had to be delivered to clients as presentations, but were assembled manually. Static templates could not adapt to varying data, while unconstrained LLM output produced inconsistent formats and uncontrolled numbers (hallucination).
Analysts focus on interpretation instead of writing reports. Reports are generated automatically and consistently, integrated directly with the dashboard/product, and report creation time dropped drastically from hours to minutes.
Taught in: AI Automation with n8n · Generative AI for Productivity
Production model performance degraded as data shifted, even while accuracy metrics looked stable. Delayed ground-truth labels made drift detection difficult, and schedule-based retraining did not match reality — sometimes wasteful when data was stable, sometimes too late when data changed drastically.
Models stay relevant without manual monitoring, performance degradation risk drops, model updates are safer and faster, and resources are used efficiently since retraining only runs when truly needed.
Taught in: Intelligent AI Agents for Workflow Automation
The data and analytics team needed data from various public sources (e-commerce, news portals, image repositories) for market research, sentiment monitoring, and content aggregation. Manual collection was not scalable and was error-prone.
The framework cut collection time from days to hours, improved data quality through the validation layer, and supports adding new data sources with minimal code changes.
Taught in: Data Analytics & Visualization
Teaching, for me, is how I make sure Indonesia's tech industry has talent that is not just theory-aware, but production-ready.
With over 350 training sessions delivered, I focus on teaching AI Automation, Generative AI, Machine Learning, and NLP through real, project-based work.
Planning AI or data training for your team, or looking for a private mentor? Send me a message on LinkedIn with your goal and team size, and I'll reply with a program outline.