Atyab Hakeem

S

GenAI Technical Advisor

Scale AI (Unicorn),

Jun 2025 - Aug 2025 (3 months)

• Built AI knowledge intelligence tool using RAG (LangChain, LangGraph, MCP), automating Slack Q&A, knowledge transfer workflows, reducing coordinator workload by 40%, transforming queries into actionable insights. • Engineered high-quality data for SOTA models by prompt engineering, evaluating 100+ challenges across DSA, SWE, & Math, while implementing workflow optimizations to increase acceptance rate and bridge knowledge gaps. • Conducted red team assessments with adversarial prompting, enhancing LLM safety and human values alignment.

D

Software Engineer

Dassault Systèmes (Publicly Traded),

Sep 2024 - Jan 2025 (5 months)

• Engineered & integrated Generative AI ( GenAI ) services using Flask & WSGI server into Solidworks Visualize, expanding market reach by 20%, boosting user productivity by 70%, and enhancing creative freedom. • Implemented text-to-image & image-to-image models like Stable Diffusion to generate photorealistic shots, materials, & environment maps for 3D models, reducing rendering time by 70% and expanding creative assets by 120%. • Optimized parameters & engineered prompts, boosting quality by 20% while cutting processing time by 35%. • Created API documentation, workflows, and Figma mock UI to streamline AI integration, reducing onboarding time.

P

Personal Health Assistant Chatbot

Jun 2024 - Aug 2024 (3 months)

• Built a medical assistant using supervised fine-tuning (SFT) with LoRA on LLaMA3 with Retrieval Augmented Generation (RAG) & Vector Database (Pinecone), achieving a BERT score of 84%, improving response accuracy.

N

Graduate Research Assistant

Northeastern University,

Oct 2023 - Sep 2024 (1 year)

• Applied computer vision & NLP to 80,000+ video ads, boosting advertisement intent identification accuracy by 33%. • Improved demographic identification accuracy by 5% by experimenting with diverse algorithms such as FairFace. • Deployed video analysis pipelines on AWS EC2 and High Performance Computing Clusters (Discovery), optimizing GPU/CPU configurations to enhance processing speed by 50% and reduce operational costs by 25%. • Benchmarked LLMs like Video Llama & OpenAI GPT for advanced multimodal analysis to evaluate ad content.

L

Data Scientist

Lincode Labs (Startup),

Jan 2023 - Jun 2023 (6 months)

• Built Edge Plus, an application for the entire ML workflow, from data collection to deployment and inference. • Utilized data scraping, collection, pre-processing, and feature selection techniques to curate high-caliber datasets. • Developed machine learning detection models using PyTorch for the visual inspection of automobile parts, reducing false positives by 20%, & containerized services like OCR using Docker, dropping processing time by 30%. • Collaborated with cross-functional teams for technical support and delivered end-to-end solutions to clients.

L

Technical Mentor

Lincode Labs

About

Detail

Timeline

Résumé