Agentic enterprise platforms
Designs LangGraph orchestrators, specialized agents, tool execution, shared memory, reusable skills, and MCP connectors for private enterprise AI systems connected to scoped business data and workflows.
Full-Stack AI Engineer
Full-stack AI engineer with 4+ years building production AI systems, agent workflows, RAG, and Python services. Leads a private enterprise AI platform with shared memory, connectors, artifacts, and permission-aware retrieval. Experience includes LangGraph, MCP, FastAPI streaming, evaluation pipelines, fine-tuning, legal technology, and cloud deployment.
Lahore, Pakistan
Designs LangGraph orchestrators, specialized agents, tool execution, shared memory, reusable skills, and MCP connectors for private enterprise AI systems connected to scoped business data and workflows.
Builds permission-aware RAG, hybrid search, chunking, reranking, vector retrieval, and evaluation pipelines to measure faithfulness and retrieval quality while preserving document access permissions in enterprise systems.
Develops streaming Python APIs, background services, voice pipelines, and web interfaces using FastAPI, Flask, Django REST Framework, Celery, Redis, React, and cloud deployment tooling for production applications.
Fine-tunes and optimizes open-source models with LoRA, QLoRA, PEFT, and quantization, while deploying services through Azure, AWS, Docker, and CI/CD with practical production monitoring and evaluation.
Leads a private enterprise AI platform serving multiple departments, owning architecture and a two-engineer team while delivering multi-agent features, scoped connectors, permission-aware retrieval, streaming services, voice capability, and evaluation infrastructure.
Improved a legal-technology platform handling TRO cases by fixing production issues, integrating Gemini through Vertex AI, and refactoring asynchronous execution with threading and multiprocessing to improve throughput for time-sensitive legal workflows.
Built a live legal-services backend with billing and authentication, deployed AI microservices, and developed agent and voice workflows using LangGraph and Vapi while supporting model training and AWS-based production delivery pipelines.
Benchmarked retrieval architectures, fine-tuned open-source models, supported vector search, and built recruitment and ML applications, combining FastAPI backends with React, Streamlit, and Gradio interfaces for practical AI system delivery and demonstrations.
Delivered Python automation, backend, and data pipeline projects for freelance clients while improving KYC machine-learning pipelines at Programmers Force, supporting identity verification for a fintech banking product and asynchronous client delivery workflows.
Highlights from professional projects, product delivery, and specialist work.
Built a private AI workspace with multi-agent chat, shared project memory, skills, artifacts, and scoped MCP connectors. Implemented LangGraph routing, FastAPI token streaming, Celery jobs, and evaluation while leading architecture and a two-engineer delivery team.
Engineered a 16TB pipeline from SharePoint to Azure AI Search with security-trimmed retrieval. Connected document access permissions to AI answers, benchmarked retrieval behavior, and used RAGAS and LangSmith to measure quality and faithfulness before release.
Compressed Falcon and LLaMA 2 using four-bit quantization and LoRA adapters, then benchmarked latency, memory use, and accuracy. The work examined practical trade-offs for low-resource inference alongside hands-on Python implementation and open-source model experimentation.
University of Engineering and Technology, Lahore
Sep 2018 – Jun 2022
Builds private AI platforms that connect multiple departments and business systems while preserving document permissions, scoped tool access, shared context, and measurable retrieval quality.
Owns agent orchestration, streaming APIs, connectors, evaluation, cloud delivery, and stakeholder scoping, combining practical full-stack engineering with retrieval research and model optimization experience.
Share your requirements, current challenges, and delivery goals. Botmer will review the brief and help arrange the next step.