Data Science Lab
Ripley Retail IQ
Multichannel analytics platform for Ripley that consolidates sales, inventory and demand intelligence into a single console for commercial and logistics decision-makers.
AI Engineering · Latin America
We are Forcast. End-to-end AI engineering, built to fit your needs.
We design, build and operate AI systems for companies in Latin America. Proprietary technology, MLOps methodology and measurable results in production.
300+projects
60+clients
8+years
4countries
Technology partners
Google CloudAmazon Web ServicesMicrosoftAIClaudeOpenAI
They trust us
What we do
End-to-end services
Three AI fronts that move the needle, plus classic engineering to take them to production.
AI · Core layerWhere we make the most impact
01 / 03
AI Agents & Orchestration
We design and build intelligent AI agents that automate and orchestrate your processes
- Multi-agent systems
- LLM orchestration
- RAG pipelines
02 / 03
Token Optimization
Maximize your AI's performance and minimize costs with intelligent token optimization
- Advanced prompt engineering
- Cost reduction
- Token consumption analysis
03 / 03
AI Infrastructure Security
We protect your AI infrastructure against vulnerabilities, attacks and data leaks
- Prompt injection defense
- Data privacy & PII
- Model security
Engineering and DataFoundations for taking AI to production
Data Science Lab
Answer questions and make predictions with algorithms
Image Processing Suite
A camera intelligence platform that identifies whatever you want
Software Factory
We build custom, flexible and easy-to-use products
Blockchain
We integrate and develop systems for your processes and services on the blockchain
About Us
Our team
A team of engineers, data scientists, GenAI specialists and leaders building end-to-end AI for companies in Latin America.
Oscar Castillo
Co-founder & Builder
San Mateo · Santiago
Entrepreneur and builder with 12+ years building companies in AI infrastructure, developer tools and computer vision.
View profile
Rodolfo García García
Head of Sales & Marketing Strategy
Business engineer with a Master's in Marketing. He has led commercial strategy, go-to-market and digital projects at Forcast for more than 6 years.
View profile
Francisco Muñoz Ponce
Data Scientist & Software Engineer · GenAI
Data Scientist and Software Engineer specializing in GenAI, computer vision and MLOps. Master's in Computer Science from the Universidad de Chile.
View profile
Rosa Berrú
Commercial Manager
Executive specializing in talent management and innovation, focused on data- and AI-driven technology projects.
View profile
Isabela Meza Morales
Graphic Designer · Brand Identity
Graphic designer with 3 years of experience, specializing in brand design, visual identity and UX/UI.
View profile
Rubén Román Vilasau
Software Engineer · Fullstack & Cloud
Fullstack Engineer focused on React/Next.js, Node.js and AWS. Experience in AI, blockchain and enterprise platform projects in LATAM.
View profile
Margarita Castillo
Head of Innovation · CIDIF Forcast
Business engineer with a Master's in Finance and a Diploma in Data Science. 6+ years in Business Intelligence, operations management and technology innovation projects.
View profileCases in production
What we do, done
Real platforms in operation with our clients. A representative selection — see the full catalog below.
Software Factory
Lipigas Energy Ops
Operational monitoring platform for Lipigas that centralizes the status of plants, fleets and distribution with real-time alerts and executive dashboards.
Data Science Lab
ACOT — Wildfire Prediction
Desktop platform that uses AI and machine learning to predict wildfire spread patterns, with interactive maps and live configurable variables.
Data Science Lab
MOVIA — Antu
Custom platform for Antu that manages the monitoring of electric vehicles and chargers with data science intelligence, delivering charging optimization strategies to reduce energy costs.
Software Factory
Tetris 4D — Logistics Optimization
Logistics optimization platform that uses AI to improve warehouse space utilization, truck load distribution and delivery routes, with an interactive 3D interface.
Image Processing Suite
Solar Tracker
Monitoring system for optimizing solar plant performance that standardizes data capture and detects panel faults, reducing maintenance costs.
Image Processing Suite
Enel — IPS Solar Plant
Image Processing Suite pipeline for automated inspection of photovoltaic panels: anomalous thermography detection, criticality ranking and traceability by serial number.
Software Factory
CCU — Quality Audits
Centralized web platform that consolidates quality audits of production plants, with automated alert generation for critical events and real-time monitoring.
Software Factory
HDI Seguros
Back-office platform for HDI Seguros that digitizes underwriting and claims workflows with configurable business rules and regulatory traceability.
Image Processing Suite
Phineal — I-V Tracer
Online I-V tracer monitoring platform for assessing the condition of individual photovoltaic panels, with real-time alerts and automated generation of quality indicators.
Data Science Lab
Instacrops — Smart Agriculture
Agricultural application that displays minute-by-minute crop station data with historical analytics, generating preventive alerts for farming decisions.
Software Factory
Elecnor — Field Reporting
Mobile reporting app for field supervisors, with offline support, activity tracking, machinery details, workforce management and photo uploads.
Software Factory
Starken — Logistics Visibility
Visibility platform for Starken operations that integrates tracking data, SLAs and branch performance with live dashboards.
Software Factory
Dober — Operations Platform
Cross-functional platform for Dober that unifies operational processes, with UX focused on distributed teams and role-configurable workflows.
AI Agents & Orchestration
Orquesta — Prompt to Production
AI agent orchestration platform: the team sends prompts from the dashboard, agents run on client machines and everything is logged. Centralized management of prompts, models and workflows with real-time token usage analytics.
AI Agents & Orchestration
Dobi Agent — IoT Web3 Assistant
Conversational agent integrated into Dobi (dobi.guru): assists operators of IoT assets (EV chargers, solar panels, smart meters, wind turbines) and orchestrates on-chain payments with DOBI tokens on the Base network. Web3 wallet auth, verified transactions and live monitoring.
AI Agents & Orchestration
Prowl — Agent Discovery Network
Middleware between AI agents and SaaS services. Agents discover, evaluate and trust APIs through a single endpoint; SaaS products get indexed, scored and certified. ASO (Agent Search Optimization) with verified benchmarks, schema drift detection and x402 micropayments.
Token Optimization
PlanckBot — Token Optimization Layer
Adaptive layer of small models that sits between the host LLM (Claude, GPT-4) and its tools. It trains one LoRA per tool (<1B params, SmolLM2 base), filters and compresses I/O in real time, and drastically reduces token consumption without losing fidelity.
Blockchain
Dobprotocol — Liquid RWA DEX
The financial layer of the Machine Economy. A real-world asset (RWA) DEX that enables fractional ownership of machines — solar panels, data centers, mining trucks — with dUSDC, on-chain oracles and the DobValidator registry. Built with DobDex, DobLink, DobValidator and Token Studio.
What our clients say
Results that speak for themselves
Verified testimonials from leaders at companies that operate with Forcast.
Forcast brought technology and artificial intelligence to the heart of our operations. They quickly understood our challenge, proposed a scoped solution and put it into production with measurable results. Today we have a long-term technical partner.
Ignacio Puga, CEO, Mirasub
The reduction in maintenance costs and a significant increase in the reliability of preventive analyses of our assets radically changed the way we carry out our procedures. Thanks to a cohesive team that was involved in every stage of the process, we can say we are pioneers in our industry.
Juan Pablo Oyanedel, CEO, Punto Solar
Thanks to the introduction of new technologies in our operations area, we have seen a significant, sustained increase in sales and many new tools that make our work easier.
Pablo Perez, CEO, Innspatial
Forcast became a technology partner that understands our business. They brought artificial intelligence into our processes in a practical way and supported us at every stage of the project, with concrete, measurable results from the very first weeks.
Horacio Contreras, CEO, AustralFalcon
Technology
Our stack
We prioritize the AI layer — then choose the infrastructure and language that best fit each project.
AI & LLM OpsCore layer — what sets us apart
OpenAILLM
Anthropic ClaudeLLM
LangChainframework
LlamaIndexRAG
Hugging Facemodels
Ollamalocal
OrquestaLLM Ops
pgvector / Pineconevector
Data Science & ML
Cloud & DevOps
Backend & Data
Frontend
Frequently asked questions
What we're always asked
How long does it take to get an AI project into production?
An honest pilot starts in 4–6 weeks. A production system with MLOps takes 10 to 16 weeks depending on complexity. We turn down projects that don't have a clear metric before starting.
Do you work with our current stack or impose your own?
Our stack is yours. We integrate with whatever you already have in production — AWS, GCP, Azure, on-prem. We only propose changes when the cost of maintaining what exists exceeds the cost of migrating.
How is the ROI of a project with Forcast measured?
Before writing a single line of code, we define 1–3 business KPIs (time, cost, quality or revenue) and their baselines. We deliver live dashboards from the first production sprint.
What happens when the project ends?
We include technical handover, operable documentation and maintenance support with a clear SLA. Ownership of the code and models is 100% the client's.
Do you work outside LATAM?
Our focus is Chile, Colombia, Mexico and Peru. We have done specific projects in the US and Spain, but always with an operational anchor in LATAM.
Let's talk
Ready to put AI to work in your operation?
Book 30 minutes with our technical team. We review your case, propose a scoped pilot with clear metrics, and you decide afterwards.