12.12.2023 aktualisiert

**** ******** ****
60 % teilweise verfügbar

Generative AI Engineer & Agile Product Manager

Berlin, Deutschland
Nur Remote
Bsc Computer Science
Berlin, Deutschland
Nur Remote
Bsc Computer Science

Profilanlagen

CV - Alexandre Ramalho

Skills

HTMLJavaScriptAPIsAgile MethodologieKünstliche IntelligenzAmazon Web ServicesAmazon S3Computer VisionMicrosoft AzureCloud ComputingContinuous IntegrationInformation EngineeringETLData WarehousingDevOpsSpeicherverwaltungAmazon DynamoDBGitHubGPSMapReducePythonPostgreSQLMongoDBMySQLObjektorientierte Software-EntwicklungOpen SourceSQLTelegrafieBestärkendes LernenPyTorchReact.jsLarge Language ModelsApache SparkGenerative AIAWS LambdaBackendKotlinpandasMatplotlibBackend-programmierungData LakeCodebaseFeature ExtractionTerraformGPTServerless ComputingGolang
AWS Lambda, Agile Coach, DynamoDB, S3, AWS, Apache Spark, API, AI, backend, Cloud, codebase, Computer Vision, CI/CD, data lake, data warehouses, DevOps, ETL, feature extraction, GPT, Generative AI, Github, GPS, Golang, HTML, Data Engineer, Computer Science, Javascript, Kotlin, LLM, Map Reduce, matplotlib, memory management, Azure, MongoDB, MySQL, OOP, open source, pandas, Postgres, PostgresSQL, Python, Pytorch, ReAct , reinforcement learning, SQL, serverless, Telegram, TerraForm

Erfahrung

Berufserfahrung5 - 10 Jahre
Freelancer seit2022

Sprachen

DeutschGrundkenntnisseEnglischverhandlungssicherPortugiesischMutterspracheSpanischgut

Projekthistorie

Lead Software Engineer

infolab.ai

< 10 Mitarbeiter

Title: Edubot - RAG Bot for University Student and Professors TECH STACK
Emp loyer: Infola b.a i
Role: Lead Software Engineer
Duration: 3 Months (9.2023 - 12.2023)

Goal:
Provide a set of RAG systems to speedup textbook and article based learning
while providing AI-enhanced analytical summaries to academic professors.

Solution:
Devised the collaborative framework for the development team, pushing
standards and methodologies to ensure rapid development cycles.
Owned the data collection procedures to drive business insights.
Maintained a Langchain Dense Passage Retrieval QA pipeline with OpenAI
Embeddings over curated document base.
Implemented state-of-the-art techniques for RAG optimization, such as query
enrichment through metadata embedding.
Deployed the dockerized application to Azure using Terraform

Lead AI Engineer

Self-Employed

URL: https://heyjarvis.co

Goal
Agent based note manager to aid in productivity, organization, and academic productivity


Solution
Created a ReAct Agent with smart-memory management using Langchain and
deployed it serverlessly via a dockerized FastAPI.
Implemented complex Data Connectors for third party applciations such
Whatsapp, Telegram, Notion, converting them into agent-usable tools.
Customly scraped and converted complex HTML structures into LLM Readable
markdown summaries of a plethora of websites, including Youtube, Github,
and others.
Experimented with locally deployed Whisper / Whisper JAX for home assistant
evaluation
Deployed the dockerized application serverlessly to AWS Cloud

Lead AI Engineer

Hippo AI Foundation

10-50 Mitarbeiter

Title: Askpaper.ai - A SOTA academic RAG system

Emp loyer: Hippo AI Foundation
Role: Lead AI Engineer
Duration: 6 Months (1.2023 - 6.2023)
URL: https://askpaper.ai

Goal
Provide a GPT powered chatbot to aid researchers extracting and agreggating
datasets information contained in academic documents



Solution
Deployed academic OCR and Computer Vision models for text and table
extra c tion
Implemented a QA Map Reduce pipeline for academic papers using Langchain
framework using HuggingFace models as backbone
Developed an accompanying Chrome Browser Extension to enhance user
experience
Orchestrated the data collection and insight generation workflow through user Chroma
research, feedback components and dashboard visualizations Vector DB

Strategically utilized a serverless AWS infrastructure, leveraging DynamoDB
and AWS Lambda for resource optimization, successfully scaling the product's
user base to over 30,000 while maintaining an exceptional 99.99% uptime

Full Stack & Data Engineer

Volkswagen AG

>10.000 Mitarbeiter

Goal
Transitioning from several scattered on premise data warehouses to a Cloud
based centralized data lake, consolidating 24TB worth of data and enabling
advanced analytics across the 12 Volkswagen Group brands procurement data.


Solution
Succesfully designed and developed an orchestrated backend to ingest and
transform the data.
implemented over 500 stories in a 100k+ LoC codebase, which included a
PostgresSQL DB and fully automated Python Apache Spark ETL Pipeline,
hosted on AWS (using RDS, Glue, Athena, S3)
Ensured continuous stability, reliability and replicability through extensive unit
and end-to-end testing, Infrastructure as Code and CI/CD.
Served as Agile Coach for 3 teams for a total of 12 developers for over a year.

Portfolio


Kontaktanfrage

Einloggen & anfragen.

Das Kontaktformular ist nur für eingeloggte Nutzer verfügbar.

RegistrierenAnmelden