a
abtaal_ai

Abtaal Aatif

@abtaal_ai

AI Engineer

Finlândia
Inglês, Urdu
Algumas informações são exibidas no idioma inglês.
Sobre mim
I am an AI Engineer based in Helsinki with over a year of experience across enterprise AI, computer vision, and LLM agent systems. I have built multimodal RAG pipelines at IBM and shipped LLM agents into enterprise products at Data Science Dojo. I specialize in PyTorch, FastAPI, and agentic workflows.... Saiba mais

Habilidades

a
abtaal_ai
Abtaal Aatif
offline • 
Tempo médio de resposta: 1 hora

Conheça meus serviços

Consultoria de Tecnologia de IA
I will audit your ai chatbot and fix wrong answers
A partir deUS$ 45 /projetoMais detalhes

Portfólio

Experiência profissional

IBM

AI Engineer

IBM • Período integral

Nov 2025 - Present • 10 mos

Built a multimodal RAG system using FastAPI and vision-language models for infrastructure maintenance, achieving 98% top-1 accuracy. Developed a React Native app for real-time asset identification via YOLOv11 and ArcGIS. Fine-tuned YOLOv11 for pothole detection and analyzed IoT sensor data for predictive maintenance.

Data_Science Dojo

Associate Data Scientist II

Data Science Dojo • Período integral

Feb 2025 - Nov 2025 • 9 mos

Built and shipped AI agents into two enterprise products used by real customers. Delivered 15+ agents using LLM orchestration, with several demonstrated at industry conferences. Connected 6 external tools (Slack, HubSpot, Azure DevOps, Algolia, Perplexity, Code Runner) through MCP servers running on Docker and Azure, so users could work across all of them in plain language. Improved agent tool-routing accuracy from 65% to 85% by analysing over 1,000 real conversations to find where agents picked the wrong tool, then rewriting the prompts controlling routing and execution order. Also built automated alerting across 12+ projects, connecting error tracking to the team's ticketing and CRM systems.

Plutus21_Capital

AI Research and Development Intern

Plutus21 Capital

Oct 2024 - Feb 2025 • 4 mos

Built an AutoGen multi-agent system using Llama to map company names to stock tickers and generate yfinance code for automated financial data retrieval and analysis.