Help & Supportcareers@gravityer.com
Generative AI Engineer with 2+ years in LLM Systems & Production Deployment
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Production-focused Generative AI Engineer who architects, deploys, and operates LLM-powered systems at commercial scale. Engineered a Python/Flask LLM email automation API (Groq) handling 50+ daily B2B requests with 80% reduction in manual workload; a full-duplex WebSocket voice agent achieving sub-200ms end-to-end latency; and a multilingual conversational AI backend processing English and Urdu via WhatsApp. Building AURA OS - a multi-agent autonomous AI system shortlisted for the Google AI Seekho Hackathon. Deploys exclusively to production on AWS and Docker. Thinks in LLM pipelines, ships in production.
University of Lahore
BS Embedded Systems · Embedded Systems
August 1, 2024 – June 30, 2028
Punjab Group of Colleges
Intermediate · Pre-Medical
June 1, 2022 – May 31, 2024
Afrium
Founding Campus Ambassador
June 1, 2026 – Present
Pakistan
Fiverr (manaal_ai)
AI Automation Engineer
April 1, 2026 – Present
Pakistan
DecodeLabs
Artificial Intelligence Intern
April 1, 2026 – May 1, 2026
Remote
Pakistan Health Parliament
Head of Digital Media
February 1, 2026 – Present
Lahore, Punjab, Pakistan
Lojics
Business Development Executive
January 1, 2026 – March 1, 2026
United Kingdom
DS Trainings
WordPress Developer
July 1, 2025 – December 1, 2025
Lahore, Punjab, Pakistan
Brain Creatives
Digital Presence Manager
March 1, 2025 – June 1, 2025
Lahore, Punjab, Pakistan
enARtifi
Associate Digital Presence Manager
February 1, 2025 – June 1, 2025
Lahore, Punjab, Pakistan
ZUNF Medicare / PITB Incubation Wing / HiSkyTech
Graphics Designer & Digital Marketer
July 1, 2024 – December 1, 2024
Lahore, Punjab, Pakistan
AURA OS - Multi-Agent Generative AI System
June 1, 2026 – Present
Autonomous multi-agent AI operating system that transforms unstructured inputs into coordinated LLM-driven actions under real-world constraints. Shortlisted for the Google AI Seekho Hackathon as the flagship advanced agentic AI project. Architected a multi-agent orchestration layer with event-driven coordination → specialized LLM agents own discrete operational domains and communicate via a shared event bus, eliminating single-agent bottlenecks. Engineered structured-output LLM reasoning pipelines with grounded, auditable decision boundaries using Groq; prevents hallucinated actions in mission-critical agentic workflows. Built real-time operational state machines that detect anomalies and autonomously re-prioritize tasks without human intervention → deployed with full observability and recovery logging. Implemented graceful degradation architecture: agents fail independently without cascading failures; autonomous recovery is logged and observable on AWS-hosted infrastructure.
LLM Email Automation API
June 1, 2026 – Present
Automated triage and response system for 50+ daily inbound B2B emails → eliminating manual classification and scaling team throughput without headcount increases. Built a Python Flask REST API routing inbound requests through Groq LLM: intent classification → context-aware response generation → Gmail API dispatch → Google Sheets CRM persistence → full AI workflow automation. Dockerized and deployed on Railway with a Streamlit operator dashboard; commercially distributed via Gumroad with Zapier/Make webhook templates for no-code client onboarding. Delivered 80% reduction in manual email handling; system operates autonomously with zero ongoing engineering intervention → demonstrating production-grade AI deployment at scale.
Real-Time AI Voice Agent (Priya)
June 1, 2026 – Present
Low-latency conversational AI calling system engineered for full-duplex real-time voice dialogue – eliminating HTTP polling overhead. Architected a full-duplex WebSocket pipeline: audio in → Deepgram STT → Groq LLM inference → TTS synthesis → audio out, achieving sub-200ms end-to-end latency in live production operation. Engineered persistent session state management across conversation turns; structured entity extraction writes to CRM automatically via REST API – fully integrated conversational AI workflow. Deployed with interruption handling, dynamic follow-up generation, and a self-serve analytics dashboard → production-grade AI deployment on cloud infrastructure.
Multilingual WhatsApp Conversational AI Backend
June 1, 2026 – Present
Engineered a webhook handler pipeline: Deepgram STT voice transcription → Groq LLM response generation → structured data persistence, supporting 24/7 multilingual consumer handling in English and Urdu. Designed discrete microservice layers enabling independent component swapping without full redeployment → reduced response latency from hours to seconds with full language and LLM model extensibility.
LensIQ AI Production Computer Vision Pipeline
June 1, 2026 – Present
Engineered a multi-pipeline AI vision system: YOLO object detection with confidence scoring → OCR text extraction → LLM-powered automated scene summarization → annotated visualization – full detection-to-insight pipeline. Deployed an interactive visual question-answering module enabling natural-language image queries, allowing non-technical users to interact with CV inference directly via Streamlit.
Edge AI Surveillance Platform
June 1, 2026 – Present
AI-powered ESP32-CAM surveillance with motion detection, face recognition, and Firebase Realtime Database integration; paired with Google Voice + MIT App Inventor mobile app for live alerts and remote monitoring; model-swappable without reflashing.
ESP32 IoT Occupancy System
June 1, 2026 – Present
Ultrasonic sensor fusion with interrupt-driven firmware; live telemetry streamed to ThingSpeak cloud dashboard; schema designed for zone expansion without firmware changes.
Certiport Cloud Specialist
Microsoft / Cloud Fundamentals
June 1, 2026 – Present
Programming Fundamentals
Duke University & UC Santa Cruz
June 1, 2026 – Present
AWS Student Community Day - National Delegate
Enterprise Cloud Forum
June 1, 2026 – Present
Raspberry Pi & Python Programming
University of California, Irvine
June 1, 2026 – Present
AI Connect 2026
University of Lahore
June 1, 2026 – Present
Cultural Fit Analysis
The candidate's diverse project portfolio, ranging from real-time voice agents to multi-agent systems and computer vision pipelines, indicates a broad interest and adaptability within the AI domain. Their freelance work and involvement in community initiatives suggest a proactive, self-starter mentality. The focus on building commercialized and production-grade solutions aligns well with a results-driven culture. Their engagement in various roles, including business development and digital media, alongside technical roles, suggests a well-rounded individual capable of understanding broader business contexts, which is valuable in cross-functional teams.
Soft Skills & Operational Fit
The candidate demonstrates strong initiative, a proactive approach to learning, and a clear entrepreneurial spirit through their freelance work and participation in various community and leadership roles. Their project descriptions highlight an ability to work autonomously, manage complex systems, and focus on production-grade deployments, suggesting a good operational fit for roles requiring self-direction and a results-oriented mindset. The emphasis on real-world constraints, observability, and graceful degradation in projects like AURA OS indicates a mature approach to system design and reliability.