Wasiq Malik Email & Phone Number

New York, Ny, Us

Lahore, Pk

New York, New York, United States

Lahore, Punjab, Pakistan

Leading a team of 4 engineers, delivering LLM apps to our platform to supercharge learning experiences using the OpenAI API with GPT-4 and fine-tuning GPT-3.5 on our private data.Led the development of a semantic code search engine over our codebase using GPT-4 Turbo, hugging face embeddings and elastic search, reducing the time for our new engineers to be onboarded on a 2+ million LoC repository.

Lahore, Punjab, Pakistan

Shipped AI-powered search experience on devpath.com by implementing RAG using Langchain with GPT-3/4 running over a semantic search engine built using ElasticSearch in Python.Led the development of AI Chatbot Assistants powered by OpenAI's Function Calling API for product navigation and insights in both B2B and B2C, resulting in an increase of annual subs by 20%,Hosting 30+ technical interviews for hiring junior to mid-level software engineers and mentoring our new hires in the AI/ML and Code Executions domain. Increasing our engineering headcount by 30% over the last year.

Lahore, Punjab, Pakistan

Led the product and engineering initiative for generative AI integrations, and developed a code feedback module built on the ChatGPT API and Langchain. Enabling educative.io to be the first e-learning platform offering AI-powered personalized code reviews in the market. Resulting in the addition of over 250 DAU.Led the product initiative and development for Sandpack integration using TypeScript React to enable in-browser compilation for frontend technologies (react.js, vue.js, angular.js), resulting in over 50% faster executions and 25% less compute costs. Led multiple team members in smaller initiatives involving caching Sass compilation via Cloudflare KV workers with a 90% hit rate and optimizing the cold start of our Code Compilation micro-service by 50%.Developed features for our high-impact Leetcode-style coding playground, CodeTest (averaging 10 executions/min) by implementing a code evaluator and DSA in languages like Python, JavaScript, C++, Java, Go, and Rust.Added IntelliSense and breakpoint debugging support for 10+ programming languages in our live workspace (IDE) by integrating OpenVSCode Server into a NodeJS micro-service, improving UI/UX for our learners.Improved our Firecracker microVMs launch time by over 82% (down to 1s) by implementing kernel-level full-memory snapshotting in Golang using the Firecracker Go SDK.

Lahore, Punjab, Pakistan

Added nested virtualization support for firecracker VMs by building a custom Linux kernel OCI image with KVM-enabled features, resulting in expanded capabilities for learners to utilize technologies like Docker Swarm, Kubernetes, and LXD. Developed a full-stack solution for testing learner container state in our live coding environments using TypeScript React, Python Flask API, and Google Cloud Firestore. This enabled the unguided projects vertical in our product offerings. Optimized our live coding environments' startup time by up to 75% on average by designing a scalable caching system for our Docker images using an NFS share over Google Cloud Filestore, enhancing user experience and productivity.Migrated our Cloudflare infrastructure to HCL code using Terraform, reducing manual config and production bugs by 20%. Enhanced the performance of our Android App Development Widget by integrating AOSP Cuttlefish Emulator. This optimization led to an 83% improvement in launch times and reduced GCP compute costs by 85%.

Bellevue, Washington, United States

Building an interactive text-based course to introduce software developers and engineers to the world of Quantum Computing with a focus on quantum algorithms and their time complexity analysis in comparison to classical algorithms.

Lahore, Punjab, Pakistan

• R&D for a novel deep-learning solution to the problem of real-time identification of soccer players inlive/broadcast videos.• Created an In-house Soccer players dataset extracted from Premier League and LaLiga videos.• Developed a real-time player tracking pipeline in GPU-accelerated PyTorch using DeepSort withYOLOv5 backend, reaching over 85% accuracy.• Implemented Jersey number detection using YOLOv5, fine-tuned on our Soccer Player Dataset,achieving over 90% accuracy.• Implemented Jersey number recognition using ResNet trained on SVHN and fine-tuned on ourDataset, achieving 87% accuracy.

Lahore, Punjab, Pakistan

• Fine-tuned GPT-2 to generate full-length text articles based on an input prompt and its semantics.• Built training/inference pipelines using huggingface transformers and a REST API using cortex.dev• Deployed on an AWS EKS cluster with V100 GPUs, resulting in a 2-4s inference time.• Provided support for production bugs and maintenance after go-live for writeme.ai

Dayton, Ohio, United States

• Applying Sentiment Analysis on text using LSTMs, K-Means, and roBERTa, using the results tocompare LSTMs and Transformer models in terms of inference/training speed and task accuracy.• Utilized trained models to analyze tweets, achieving over 85% accuracy.

Bachelor Of Science - Bs, Computer Science, 3.79/4.0

• Magna Cum Laude • Gold Medal - Fall 2020 • Dean's List • Rector's List

Aws Machine Learning Engineer Nanodegree, Computer Science

Intro to ML Developing ML workflows Deep Learning for computer vision and NLP Operationalizing ML on AWS Sagemaker Build ML Portfolio

Intro To Quantum Computing Certification, Quantum Computing

In this year-long course, over 7,500 students from 125 countries learned skills in quantum mechanics, quantum computation, quantum.

Certificate Of Quantum Excellence, Quantum Computing, Quantum Machine Learning, 82/100

Completed the two-week intensive course provided by IBM Quantum, completing all graded lab work assignments with a final cumulative score.