a
alii_shahzad

Ali S

@alii_shahzad

Software Engineer

Pakistan
Engels, Urdu
Sommige informatie wordt in het Engels weergegeven.
Over mij
I am a software engineer with experience in developing AI-driven tools and API integrations. I have worked on projects involving machine learning, natural language processing, and audio processing. I am skilled in using technologies like FastAPI, Qdrant, Langraph, and Azure AI.... Lees meer

Skills

a
alii_shahzad
Ali S
offline • 
Gemiddelde reactietijd: 75 uur

Bekijk mijn diensten

AI-integraties
I will build autonomous multi agent systems using langgraph, langchain
VanafUS$ 90 / projectMeer informatie
Automatiseringen en agents
I will develop ai voice agents with vapi, livekit, fastapi
VanafUS$ 50 / projectMeer informatie

Portfolio

Werkervaring

Software Engineer

APIMatic.io • Fulltime

May 2023 - Present • 3 yrs

Engineered AI-driven Model Context Protocol (MCP) tools, servers, and chatbot context plugins for enterprise clients including PayPal and Shell, achieving one-shot SDK integration results with modern coding assistants. Developed the OpenAPI Specification (OAS) importer alongside APIMatic's VS Code extension and GitHub application to streamline the validation and importing of API specifications.

Machine Learning Engineer

108ai • Fulltime

Nov 2022 - May 2023 • 6 mos

Developed an intelligent legal AI assistant leveraging Qdrant, Langraph, and Azure AI for RAG workflows. Implemented FastAPI for scalable API development and improving legal document processing. Worked on integrating LLMs for advanced document comprehension and retrieval and implemented chunking for Legal Documents.

Machine Learning Engineer

Rapidev NSTP • Parttime

May 2022 - Aug 2022 • 3 mos

Built a Transformer-based suspect matching system integrating audio enhancement, source separation, and speaker verification with Wav2Vec 2.0 and SpeechBrain. Designed an optimized audio pipeline for ASR and speaker identification using log-mel spectrograms, MFCCs, and adaptive preprocessing techniques. Enhanced model accuracy by fine-tuning self-supervised embeddings and applying SpecAugment, noise reduction, and real-world dataset adaptations.