Thiago Silva

Senior

AI Engineer

Who I am?


Architecting the Future of Enterprise AI

I am a Senior AI Engineer dedicated to designing and deploying highly scalable, LLM-driven solutions. With a deep foundation in Natural Language Processing (NLP) and Computer Vision, my expertise spans developing intelligent agentic workflows, building internal AI accelerators, and optimizing transformer architectures for maximum performance.

My Vision

I believe the next frontier of artificial intelligence relies on Responsible AI. Building powerful systems means building secure systems. I advocate for rigorous data privacy guardrails, ethical content moderation, and the strategic use of local model deployments to give enterprises full control over their data. Beyond text and data, I am deeply passionate about multimodal AI; experimenting with audio and video generation to create the next generation of highly engaging, interactive digital experiences.

Global Collaboration

As a bilingual (Portuguese/English) professional, I thrive on global collaboration and am driven by the challenge of translating cutting-edge Generative AI research into tangible, real-world value.

Core Skills

Generative AI & LLMs

Building scalable and secure AI assistants and chatbots.

I design intelligent, multi-agent workflows that can autonomously solve complex enterprise tasks, prioritize user security, and deliver highly accurate information.

NLP, Embeddings & Vector Search

Transforming massive amounts of unstructured text into searchable, actionable knowledge.

I utilize vector databases and advanced embedding techniques to power lightning-fast, context-aware semantic search, enabling AI systems to deeply understand human language and deliver highly accurate, contextual responses.


Multimodal Generation

Architecting pipelines for the creation of synthetic audio, images, and videos. I build intelligent orchestrators capable of seamlessly combining different generative solutions into cohesive, highly engaging multimedia experiences.

Engineering & Architecture

Designing robust, high-performance architectures and microservices capable of supporting massive user concurrency. I build the essential APIs and automated data pipelines that seamlessly connect complex business logic with advanced AI models.

Cloud Infrastructure & DevOps

Deploying reliable, production-ready AI models into scalable cloud-based environments.

I implement automated continuous integration and delivery (CI/CD) pipelines to streamline updates, ensuring enterprise-grade stability and optimized performance.

Agile Delivery & Engineering Excellence

Driving the smooth execution of complex AI projects through Agile methodologies and structured task management via Jira.

I prioritize clean coding practices and comprehensive technical documentation, while fostering open, transparent communication across cross-functional teams to ensure alignment, continuous collaboration, and successful product delivery.

Projects

Recent projects that I've been working on

a person holding a smart phone and a credit card
Secure Financial AI Assistant

A customer-facing AI agent designed to securely handle sensitive financial data and user inquiries.

Developed for a global financial services client, this solution implements rigorous Responsible AI guardrails to safely process complex account information and securely communicate benefits to users while maintaining enterprise-grade privacy.

a close up of a cell phone with icons on it
Agentic AI Evaluation Framework

An advanced benchmarking engine used to test, standardize, and validate AI agents across diverse enterprise industries.

Built utilizing synthetic data and research-backed personas to rigorously evaluate agent orchestration. The framework uses custom rubrics to score tool selection, reasoning, and overall response quality against established ground-truth data.

Healthcare AI Workflow

A highly reliable, bilingual medical AI assistant engineered to support over 600,000 concurrent users.

Engineered a specialized multi-agent workflow designed for the healthcare sector. The system handles intelligent medical query routing, real-time PII redaction, geo-location matching for nearby services, and robust content moderation.

Tech Stack

A full breakdown of technologies, services, and libraries that I have experience with

Generative AI

Focusing on state-of-the-art implementation and orchestration.

Frameworks: LangChain, LlamaIndex, LangGraph, Langflow, CrewAI.

Model Providers: OpenAI (Since GPT-3), Anthropic, Google (Gemini), Mistral, Meta (Since Llama 2), AWS Bedrock.

Generative Media: ComfyUI (Custom Node Workflows), Stable Diffusion (XL/Flux.1), Midjourney, Google Veo, Runway Gen-3.

Techniques: RAG (Retrieval-Augmented Generation), Fine-tuning (LoRA), Prompt Engineering, Vision-Language Models (VLM).

Foundations: Embeddings (OpenAI, HuggingFace)

Responsible AI

Engineering defensive AI architectures, ensuring ethical governance through red teaming and automated guardrails.

Threat Mitigation: Prompt Injection (Direct & Indirect), Payload Splitting, Jailbreaking defense.

Red Teaming: Adversarial Testing, LLM Vulnerability Scanning, OWASP Top 10 for LLMs.

Guardrails: AWS Bedrock Guardrails, LangChain Guardrails, Data Anonymization, Content Filtering.

AI Observability & Evaluation

To monitor model performance, hallucination rates, and system latency.

Evaluation: LLM-as-a-Judge (G-Eval), DAG, QAG.

Benchmarking: Agentic Trajectory Accuracy, Hallucination Detection.

Tracing & Logging: Langfuse, LangSmith.

LMOps: Cost/Latency Analytics, Semantic Caching.

Data Engineering & ETL

Building high-throughput data pipelines and automated ETL processes

Orchestration: AWS Glue, Airflow, Argo Workflows.

Extraction: Scrapy, Selenium, Beautiful Soup.

Data Modeling, ETL Pipelines, Large Scale Datasets.

Databases & Vector Search

Designing hybrid storage strategies that combine traditional relational data with optimized vector search.

Storage: PostgreSQL, MongoDB, MySQL, Microsoft SQL Server.

Vector & Graph: ChromaDB, Neo4j, Amazon OpenSearch.

Cloud Storage: AWS S3, Azure Blob Storage, Google Cloud.

Computer Vision & ML

Extracting actionable insights from visual data through traditional image manipulation, custom OCR pipelines, and modern object detection architectures.


Vision: OpenCV, Detectron2, Tesseract OCR, GPT-Vision.

Frameworks: PyTorch, TensorFlow, Keras, Scikit-learn.

Processing: Video file manipulation, Pillow, NumPy, Word2Vec.

Cloud & Infrastructure

Managing scalable workloads

AWS Ecosystem: Amazon EC2, Bedrock, Lambda, S3.

Compute: Kubernetes, Docker, Linux.

Architecture: Microservices, Data Lakes.

MLOps & Deployment

Designing hybrid storage strategies that combine traditional relational data with optimized vector search.

Serving: FastAPI, Flask, Streamlit.

CI/CD: GitHub Actions , Argo Workflows , Canary, Version Control.

Messaging: RabbitMQ , Celery.

Software & Automation

Applying expert-level software engineering standards

Languages: Python, SQL, Batch Processing.

Tooling: SQLAlchemy, AsyncIO, Requests, PyQt5.

Quality: Pytest , Documentation, Poetry/Conda.

Contact


LinkedInWhatsAppEmail