βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β ASGHAR QAMBER RIZVI Β· AI/ML ENGINEER β
β Building systems that think, remember, speak β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
Python AI engineer who builds end-to-end intelligent systems β from fine-tuning LLMs and designing RAG pipelines to deploying production FastAPI backends that handle real users. I care about things that actually work: low latency, correct retrieval, reliable APIs, and models that behave in production, not just in notebooks.
Currently finishing CS @ Bahria University (3.81 GPA, graduating June 2026). Based in Karachi, Pakistan. Available for remote roles.
Python Β· FastAPI Β· Gemini AI Β· PostgreSQL/PostGIS Β· JWT Β· Render Β· Supabase
An agentic platform for Pakistan's informal economy. Users describe a service need in natural language (Urdu, Roman Urdu, or English) β JUVO's multi-agent pipeline extracts intent, finds geographically nearby providers via PostGIS spatial queries, and creates a confirmed booking, all within a single conversational flow.
- Designed a multi-agent architecture: Intent Agent β Discovery Agent β Booking Service
- Built Hold-to-Lock (HTL) reservation system with 5-minute expiry and ACID database triggers to eliminate 100% of double-booking conflicts
- Implemented role-based JWT auth (user + provider), rate limiting, background task cleanup, and full Swagger documentation
- Deployed on Render + Supabase with Docker; full provider dashboard with analytics
Python Β· LangChain Β· Llama-3 (3B) Β· LoRA Β· MongoDB Vector Atlas Β· ChromaDB Β· RAG
Domain-specific legal AI assistant built on fine-tuned Llama-3 with a production-grade RAG pipeline.
- Fine-tuned Llama-3 (3B) using LoRA (r=32, Ξ±=64) on criminal law cases and legal statutes for improved legal reasoning
- Engineered a vector database on MongoDB Vector Atlas β 0.98 accuracy on past cases and legal statutes
- Production-level techniques: streaming responses, caching, rate limiting β query speed reduced to 200ms
- Chat and message management via LangChain memory components
Python Β· FastAPI Β· WebSocket Β· Computer Vision Β· Pose Estimation
Real-time fitness AI built during internship at Meta Frolic Labs.
- Custom computer vision pipeline with pose estimation algorithms β 98% accuracy in live exercise detection
- FastAPI backend with WebSocket integration for real-time data streaming at 0.3ms latency
- Performance analysis algorithm evaluating form quality and delivering corrective suggestions through a web interface β 40% improvement in feedback system effectiveness
Python Β· SpeechT5 Β· Transformer Β· Multi-speaker TTS
Speech synthesis system targeting the voice of Zia Mohyeddin, Pakistan's most celebrated literary narrator.
- Developed SpeechT5 transformer-based architecture with multi-speaker capabilities
- 25% improvement in naturalness over existing Urdu TTS solutions
- 92% user preference in blind listening tests for voice fidelity
Python Β· PyTorch Β· GANs Β· Transformers Β· Audio Processing
Research project: converting emotional tone in speech while preserving semantic content.
- Designed a hybrid GAN + Transformer architecture β Transformer vector embeddings carry semantic context; GAN learns the target emotion tone
- Achieved 62% success rate on neutral β happy emotion conversion
- Custom approach combining contextual sentence embeddings with adversarial training for audio emotion transfer
Python Β· Transformers Β· Fine-tuning Β· NLP
Fine-tuned a language model specifically for high-quality paraphrase generation β preserving meaning while altering structure and vocabulary. Trained and evaluated on custom paraphrase datasets.
LLM & AI Llama-3 Β· SpeechT5 Β· LoRA fine-tuning Β· GANs Β· Transformers
Prompt engineering Β· RAG pipelines Β· LangChain Β· ChromaDB
MongoDB Vector Atlas Β· Vector search Β· Embedding models
Computer Pose estimation Β· Real-time inference Β· WebSocket streaming
Vision OpenCV Β· MediaPipe
Backend FastAPI Β· Django Β· Flask Β· SQLAlchemy Β· Pydantic
JWT auth Β· Rate limiting Β· Background tasks Β· REST APIs
Databases PostgreSQL Β· PostGIS Β· MongoDB Β· Vector databases Β· Redis
MLOps & Docker Β· Render Β· Supabase Β· Alembic Β· Gunicorn Β· Nginx
Deploy Multi-user concurrency Β· Production API design
Languages Python (primary) Β· SQL Β· C++ Β· Java
| Metric | Value |
|---|---|
| Exercise detection accuracy | 98% |
| API latency (WebSocket, real-time CV) | 0.3ms |
| Legal RAG retrieval accuracy | 0.98 |
| Query speed (RAG + streaming) | 200ms |
| Urdu TTS improvement over baseline | +25% |
| Voice clone user preference (blind test) | 92% |
| Emotion conversion success rate | 62% |
| Feedback system effectiveness improvement | +40% |
| GPA | 3.81 / 4.0 |
Python AI Trainee β Meta Frolic Labs (Aug 2025 β Oct 2025) Shipped real-world AI systems under senior engineers: real-time computer vision pipeline, emotion conversion research (GAN + Transformer), and Urdu TTS with voice cloning. Production-level work with FastAPI, WebSocket, and transformer fine-tuning.