Agentic AI & RAG
Agents, document retrieval, embeddings and multi-agent workflows.
TOMÁS VALLEJO · BOGOTÁ, COLOMBIA
AI Product Builder
& Agentic Software Engineer.
I build AI products that connect data, APIs, automation and business decisions, backed by 10+ years in Data, Analytics and Business Intelligence.
Open to international roles across Applied AI, Agentic AI, Data and Analytics.
AIPRODUCT
DATAAGENTSAPIsPROOF10+years in Data & BI
60+ETL processes
35+production dashboards
37cloud containers
201tests in EIA OS
CAPABILITIES
From business question to operating system.
A hands-on combination of product thinking, backend engineering, applied AI, data and production controls.Backend & APIs
Python, FastAPI, webhooks, queues, workers and integrations.
Data & BI
SQL, ETL, semantic models, KPIs, Power BI and Tableau.
Cloud & Operations
Oracle Cloud, containers, edge deployment and health checks.
Quality & Security
Testing, RLS/RBAC, audits, observability and explicit boundaries.
FEATURED WORK
Four cases for evaluating engineering, not just interfaces.
Each card shows the problem, my contribution, the stack and concrete proof. Full dossiers preserve architecture, flow, current state and limitations.03 · 2026 LIVE DEMO
CAMPAIGN INTELLIGENCE
Alquim Ads
- Problem
- Turn campaign metrics into understandable priorities without delegating critical calculations to an LLM.
- What I built
- An executive KPI layer, deterministic rules engine, LLM-assisted analysis and scheduled reporting.
META ADSANALYTICSRULES ENGINELLM
ProofPublic demo · readiness 200 · calculation, rules and narrative separated
View technical case (ES) 02 · 2026 TECHNICAL PILOT
DOCUMENT INTELLIGENCE
AuditorIA
- Problem
- Query a scattered regulatory corpus without losing context, source linkage or expert control.
- What I built
- A RAG pipeline with ingestion, multilingual embeddings, semantic retrieval, API and LLM-assisted composition.
RAGBGE-M3APIDEEPSEEK
Proof2,767 indexed items · 16 standards · health and search verified
View technical case (ES) 09 · 2026 PRIVATE DEMO
OPERATIONAL WORKFLOW
EIA OS
- Problem
- Coordinate scope, roles, evidence and financial control without mixing data across organizations.
- What I built
- A multi-tenant full-stack product with PostgreSQL/RLS, storage, workers, exports and health endpoints.
NEXTPOSTGRESQLRLSWORKERS
Proof201/201 tests · 22 forced-RLS tables · 56 migrations
View technical case (ES) 07 · 2026 PRIVATE DEMO
MULTI-TENANT OPERATIONS
PPH Bot
- Problem
- Turn WhatsApp conversations into authorized, traceable operations separated by property.
- What I built
- A multi-tenant orchestrator with WhatsApp API, OCR, transcription, RBAC, queues and an admin back office.
WHATSAPP APIFASTAPI / OCRPOSTGRESQLREDIS / BULLMQ
Proof19 properties · 984 configured units · backend health 200
View technical case (ES) MORE SELECTED WORK
Depth across data, automation and product.
01 · 2026
INTERACTIVE COURSE
Public artifact LIVE DEMOOpen case (ES) 04 · 2015—2026Graph Engineering 101
A self-contained technical course that turns agent orchestration into a visual, interactive and critically documented learning experience.
PRODUCTION ANALYTICS
Anonymized cases LIVE DEMOOpen case (ES) 05 · 2026BI & Analytics Portfolio
An anonymized selection of BI systems built across a decade of analytics strategy, modeling, ETL and executive dashboards.
EDITORIAL AUTOMATION
Recorded run DEPLOYED BACKENDOpen case (ES) 06 · 2026IA Daily
A scheduled pipeline that turns AI signals into a curated briefing delivered through email and Telegram.
LEARNING PLATFORM
Public product LIVE DEMOOpen case (ES) 08 · 2026Alquim IA Academy
A Spanish-language editorial platform for applied AI learning through paths, guides, comparisons and cases.
CONVERSATIONAL PROTOTYPE
Demonstrable landing WORKING PROTOTYPEOpen case (ES) TienditaBot
A conversational commerce concept that turns WhatsApp into a catalog, order-taking and operating flow for small businesses.
ABOUT
The combination: data leadership, AI engineering and product delivery.
I am Tomás Vallejo, Founder of Alquim-IA. I translate business needs into metrics, metrics into workflows and workflows into dependable AI systems. I work hands-on across SQL, Python, APIs, analytics, automation and agentic architecture.