01Agentic orchestration on LangGraph
Illustrative architectureMCP Multi Tool Agent Platform
An agent platform connecting useful tools, persistent state and human approval.
Retrieva LabsAI engineering agency
AI systems.
Working together.
Retrieval, agents and automation with clear control and evidence.
From a useful idea to a system people can use.
Illustrative workflowContext to outcome
OrchestratorCoordinate the flow
Selected work
All projects
02Enterprise legal AI platform
Illustrative architectureData Greffe Justice
A legal document platform with structured search, retrieval and source context.
Engineering experience behind Retrieva Labs.
Explore the architecture, decisions and available project evidence.
01 / CAPABILITIES
Useful systems. Thoughtful engineering.
The right approach starts with your problem. Retrieva Labs brings models, data and software together around a useful outcome.
Retrieval systems
Make complex documents useful through ingestion, search and answers with source context.
Explore this capabilityAgents and tools
Connect models to useful tools with persistent state and clear human control.
Explore this capabilityAutomation and voice
Bring AI into the channels and business systems people already use.
Explore this capability
Applied machine learning
Develop models and pipelines around a defined task, data and evaluation plan.
Explore this capability
AI products and SaaS
Build the backend, interfaces and integrations that turn a model capability into a usable product.
Explore this capability02 / APPROACH
Build the capability. Keep the control.
A model is one part of a working system. Context, integration, evaluation and a clear handover make the difference.
- 01
Understand the task
Define the people, data, constraints and useful outcome.
- 02
Make the system explicit
Choose the retrieval, model and tool boundaries the task needs.
- 03
Test the important paths
Evaluate useful outputs, failures, approvals and recovery.
- 04
Ship with a clear handover
Connect the backend and document how the system is operated.
WORKING WORLDWIDEYour team. Our engineering.
Your team. Our engineering.
A shared way forward.
AI engineering for teams in the US, UK, Saudi Arabia, the Gulf and Europe, with remote collaboration worldwide.
Explore international deliveryUnited StatesUnited KingdomSaudi Arabia and the GulfEurope
Clear scope, agreed working hours, regular reviews and a documented handover. Bring the problem and we can shape the build together.
Based in Peshawar, Pakistan. Open to remote projects worldwide.THE AGENCYRetrieva Labs
Production RAG and agentic AI systems.
Peshawar, Pakistan03 / RETRIEVA LABS
An AI engineering agency. Built around the whole system.
Retrieva Labs builds retrieval systems, AI agents and software that brings them into real workflows. The work spans legal document platforms, SaaS products, voice pipelines and applied machine learning.
From architecture and prompts to the backend, testing and delivery, the agency keeps the full system in view. Useful outputs, clear approval boundaries and a documented handover shape the build.
Owned by Muhammad Mubashir. Based in Peshawar, Pakistan. The selected projects reflect the engineering experience behind the agency.
EXPERIENCE / BEHIND THE AGENCY
Built through practice.
Grounded in the fundamentals.
2021 to present
AI/ML Engineer
Independent, UpworkRetrieval systems, AI agents and language model pipelines for clients. Work spans architecture, prompts, backend integration and delivery.
2023 to 2024
AI/ML Engineer
Giant Tech SolutionsAI product development, demand forecasting models and a conversational AI system with multiple endpoints serving production traffic.
2019 to 2023
BS Computer Science
UET PeshawarAI, data science and algorithms. IBM certified data scientist.
Muhammad Mubashir’s background, qualifications and independent work.
TOOLS / CAPABILITY BREADTH
The tools behind
the working system.
From models and retrieval to backend services and integrations. The task determines which pieces belong in the architecture.
- Claude API
- Claude Agent SDK
- OpenAI API
- OpenAI GPT 4o
- LangChain
- LangGraph
- CrewAI
- Model Context Protocol
- Prompt Engineering
- LLM Evaluation
- Prompt Caching
- Pinecone
- Qdrant
- Elasticsearch
- FAISS
- MongoDB
- Hybrid Search
- Advanced RAG
- PyTorch
- TensorFlow
- Scikit learn
- Computer Vision
- NLP
- OCR and Document Processing
- Time Series Forecasting
- Python
- TypeScript
- React
- Next.js
- SQL
- FastAPI
- Flask
- Node.js
- Express.js
- Docker
- Celery
- REST APIs
- MLOps
- AWS Bedrock
- Stripe
- Twilio
- n8n
- Make.com
- Zapier
- GoHighLevel
- OpenAI Whisper
- Speech to Text
- Text to Speech
- Webhooks
More workVoice, automation, applied ML and other projects08 projects
Natural language cloud automation
Conversational AWS Infrastructure Agent
A chat driven AWS assistant on Amazon Bedrock (Claude 3.5 Sonnet) with LangChain memory, letting non technical users create, update, and manage AWS resources through natural language instead of the CLI, with mandatory field confirmation before any resource changing action executes.
Recovery support SaaS backend
AI Sponsor
Backend for a 24/7 AI sponsor chatbot for 12 step recovery programs: Claude based chat with prompt caching, Stripe subscription billing with webhooks, and a WhatsApp voice pipeline combining Whisper transcription with synthesized voice replies.
Earnings call intelligence with CrewAI
Multi Agent Financial Analysis
A CrewAI multi agent pipeline that ingests earnings call transcripts and financial PDFs, delegates summarization and structuring across specialized agents, and outputs client ready report documents.
AI therapy chatbot (MVP)
Namhee
A mental health support chatbot with mood assessment and tracking, CBT/mindfulness based guidance, session memory, and automatic risk detection that surfaces crisis helplines. Shipped production ready with documented performance targets.
CNN diagnostics for healthcare
Computer Vision for Medical Imaging
CNN based classification models for brain tumor detection and diabetic retinopathy screening from medical scans, built with TensorFlow using transfer learning.
Scanned documents, made retrievable
OCR to RAG Pipeline
A five stage pipeline (PDF, image, preprocessing, OCR, structuring) converting scanned documents into clean, vector store ready JSON for downstream RAG ingestion.
Human in the loop LinkedIn & job application agent
Content Copilot
A local first, three agent tool that drafts LinkedIn posts/comments and tailored job application cover letters, with web search grounded topic scoring. Every publish/apply action stays manual, avoiding platform ToS violations.
Conversational voice support
Voice AI Support Agent
A voice support agent that answers inbound calls, handles FAQ and order lookup requests through natural conversation, and escalates to a live agent when a request falls outside its scope.
04 / FIELD NOTES
Engineering notes from the work.
Building a Production Voice AI Pipeline: WhatsApp, Whisper, and Prompt Caching
Lessons from shipping a 24/7 WhatsApp voice pipeline with Whisper transcription, Claude generated replies, and the prompt caching setup that cut input token costs by roughly 90%.
Building a Human in the Loop Multi Agent Platform with LangGraph and MCP
How I designed an agent platform on LangGraph's StateGraph with Model Context Protocol tool access, persistent checkpointed memory, and human approval steps before sensitive actions.
RAG vs Fine tuning: When to Use Which
A practical breakdown of when RAG outperforms fine tuning and vice versa, based on real production experience building both types of systems.
05 / LET’S TALK
What are you
building next?
Bring the problem, the idea or the system you want to improve. We can work out a useful next step together.
Discuss your projectcontact@retrievalabs.com