AI Engineer | LLM Systems, RAG & AI Agents | Open-Source Contributor
Professional Summary
AI-focused software developer and open-source contributor with hands-on experience across AI/ML, deep learning, cloud-native systems, and large-scale open-source projects. Experienced in Python, Go, PyTorch, Kubernetes, LlamaIndex, LangChain, and modern AI APIs, with a strong track record of contributing to production codebases and collaborating through GitHub. Currently focused on building expertise at the intersection of AI systems, distributed infrastructure, and intelligent software engineering. Contributes to major open-source ecosystems, develops AI-driven projects, and actively works on technically demanding problems involving machine learning, reasoning systems, and cloud-native infrastructure.
Llama Index
Open Source contributor
Dec 2025 — May 2026 · 5 mos
Contributed to the LlamaIndex open-source ecosystem by implementing new features, improving reliability, expanding integrations, and strengthening test coverage. Contributions included asynchronous tool support, MCP discovery integration, property graph enhancements, custom LLM configuration, retry and workflow improvements, agent reliability fixes, and comprehensive testing. Collaborated with maintainers through code reviews while following production-grade software engineering practices across Python, AI agent frameworks, and Retrieval-Augmented Generation (RAG) infrastructure.
Air university
Bachelor of Science in Artificial Intelligence · Artificial Intelligence · GPA: 3.55
2029
No certifications added yet.
Research.ai
Jul 2026
I built a full-blown cognitive architecture for an AI agent, bridging short-term execution with long-term retention using a battle-tested four-tier memory system. Working memory sits on Redis for ultra-low latency reads and writes during active execution, while episodic vector stores, semantic knowledge graphs, and a procedural skill registry handle persistent state across distinct operational domains. The asynchronous sleep-phase consolidation pipeline acts as an offline background compiler, ingesting raw Redis streams, extracting atomic facts, parameterizing successful code workflows into reusable skills, and indexing episodic interactions. To prevent context pollution and storage inflation over time, your custom decay engine continuously recalculates memory retention scores using an adjusted forgetting curve—balancing base importance, time elapsed, and access frequency—to automatically archive or evict stale data while keeping high-value knowledge active.
NayaCapital
Apr 2026
A full-stack equity crowdfunding platform connecting founders and investors in Pakistan. Founders list startups, define milestones, and raise capital; investors browse and back opportunities with real-time fund tracking.
Celeb-Face-Recognition
Jul 2025
17-class celebrity face recognition using fine-tuned ResNet18 in PyTorch — with data augmentation, cosine LR scheduling, and stratified evaluation.
No patents or publications added yet.
Academic High Achiever Award
Air University · 2025
Immutable Record
This candidate's record is append-only and has not been edited since submission.
Record ID: UT-2026-629264