AI Engineer

Open to part-time

Bio

Muhammad H. is an AI Engineer with several years of experience in developing agentic AI systems for forecasting and strategic decision support. Muhammad's core technologies include Python, LangChain, Phidata, Agno, CrewAI, and FastAPI. He builds secure data pipelines and backend systems, utilizing cryptographic methods like Fernet for data integrity. His work involves API development and backend automation, with a focus on optimizing workflows and reducing inference costs. A notable project by Muhammad is Legal-Ease, a legal assistance application with a Flutter frontend and Flask backend, integrating Google Generative AI for dynamic consultations. He also developed an AI-powered career coach using CrewAI and Serper API, automating resume evaluations and job search processes. At New Effect, he cut LLM inference costs by 82% and reduced agent runtime from 15 minutes to 41 seconds through efficient architecture. Muhammad holds a Bachelor's in Computer Science from Bahria University Islamabad. He is well-suited for roles focused on AI system development and automation, particularly where strategic forecasting and decision support are key.

Expertise

PythonFlaskSQLGitNumPyPyTorchFast API

Video Intro

Assessments

General Project Assessment

(82/100)

Perfect on Implement LRU Cache AlgorithmMedium
Great on Understanding Rate Limiting in APIsMedium
Great on Ensuring Real-time Data Accuracy in Mobile AppMedium

Projects

Legal-Ease

Developed a Legal Assistance Application with a Flutter-based frontend, offering an intuitive user interface featuring chatbot interactions, lawyer directory, and PDF document management. The backend was built using Flask, integrating Google Generative AI (Gemini) to deliver dynamic legal consultations. The system enables users to upload legal documents in PDF format, automatically extracting and processing legal content. It provides bilingual legal guidance in English and Urdu, aligned with Pakistani law, while leveraging contextual question generation to gather case-specific information and deliver personalized legal advice.

AI-Powered Career Coach

Built an Agentic AI using CrewAI, Gemini 1.5, and Serper API to deliver personalized career guidance through autonomous agents. The system automates Resume Evaluation & Scoring, Realtime Job Search, Role Recommendations, and Weekly Learning Roadmaps. It intelligently analyzes CVs, finds the most relevant job openings from the web, recommends best-fit roles, and generates structured multi-week study plans to enhance career growth.

Experience

Junior Data Scientist
New EffectMar 2025 — Present
- Developed and optimized autonomous forecasting agents capable of analysing structured JSON data for strategic predictions.
- Designed and deployed code-executing agents to automate insight generation, directly aiding executive decision-making.
- Led root cause analysis (RCA) initiatives to uncover performance variance drivers, improving long-term process efficiency.
- Cut LLM inference costs by 82% through architectural redesign and streamlined output workflows.
- Built modular AI agent systems using LangChain, Phidata, Agno, and CrewAI for scalable, maintainable automation.
- Engineered backend automation for a parameterized chatbot, enabling GUI-independent forecasting capabilities.
- Reduced agent runtime from 15 minutes to 41 seconds through task parallelization and efficient architecture.
- Secured data pipelines using cryptographic standards like Fernet to ensure end-to-end data integrity.
- Improved accuracy and relevance of insights by refining JSON-driven data extraction and transformation workflows.
- Played a key role in integration planning, delivering insights that powered business automation and strategic pivots.
- Integrated LangGraph into modular AI agent pipelines to automate workflow orchestration for over 8 projects, reducing latency by 60% and enabling scalable, event-driven data processing over a 6-month deployment.
LangChainPhidataAgnoCrewAI