Headline

B.Tech Math & Computing @ IISc Bangalore | Ex-AI/ML Research Intern @ Ericsson | AI/ML Engineer | Applied AI, LLM Systems & Evaluation | PyTorch | Transformers & SSMs

About

I'm a B.Tech student in Mathematics and Computing at the Indian Institute of Science (IISc), Bangalore, working across AI/ML engineering and research from LLM systems and evaluation to efficient sequence modeling. As an AI/ML Research Intern at Ericsson India, I benchmarked WiMamba (a state-space-model architecture) against Transformers on AWS A10G GPUs, demonstrating linear scaling with 10.7x lower latency and 42.6x lower GPU memory. I evaluated geographic OOD transfer across five downstream wireless tasks (0.9626 composite score, 7.4-7.7 dB lower NMSE) and explored a soft-unfreezing parameter-efficient fine-tuning strategy that retained 98.5% of full fine-tuning performance with 75% of parameters frozen. Before that, I worked as a Machine Learning Engineer (AI Data Trainer) at Alignerr and a Machine Learning Engineer at micro1 authoring structured reasoning traces and training data for LLM decision-making and tool selection, and running data annotation, evaluation, and model validation workflows for frontier AI systems. I care about making AI systems measurable and reliable. I've built a trace-based LLM/RAG evaluation engine (deterministic metrics, failure taxonomy, regression testing), a verified tool-using coding agent with permission policies and test-gated completion, and from-scratch PyTorch implementations of S4D, Mamba, Mamba-2 and Mamba-3 with long-sequence benchmarks. Stack: Python, PyTorch, TensorFlow, Scikit-learn, NumPy, Pandas, Hugging Face, Docker, AWS (EC2, S3), Linux, Git. I'm currently exploring AI/ML engineering, applied AI, and research-oriented engineering roles internships and full-time in machine learning engineering, LLM/AI systems, and AI evaluation/reliability.

Location

Bengaluru

AI stack

aistudio

antigravity

claude code

codex

grok

hermes agent

Work experience

AI/ML Research Intern at Ericsson, May 2026 - Jul 2026 · 3 mos, Bengaluru, Karnataka, India, Researched and benchmarked WiMamba (a Bidirectional Mamba State-Space Model) against Transformer baselines for Wireless Foundation Models on massive MIMO Channel State Information (CSI). Achieved a 10.7x reduction in inference latency (30 ms vs. 327 ms) and 42.6x lower peak GPU memory usage (113 MB vs. 4.8 GB) at 4x4 patch granularity on AWS A10G GPUs. Evaluated geographic out-of-distribution (OOD) generalization across Florida->San Diego DeepMIMO datasets, achieving a 0.9626 composite score across 5 downstream wireless tasks. Designed a Soft Unfreezing parameter-efficient fine-tuning (PEFT) strategy that preserved up to 98.5% of full fine-tuning performance while cutting trainable backbone parameters by 75% for edge deployment.

SciTech Coordinator at Rhapsody, IISc, Jan 2026 - Apr 2026 · 4 mos, India, Organized and managed technical events and hosted Competitive Coding Contests, fostering problem-solving, programming skills, and student participation in coding competitions.

Machine Learning Engineer (AI Data Trainer) at Alignerr, May 2025 - Aug 2025 · 4 mos, Remote, Authored structured reasoning traces and high-quality training data to support LLM decision-making, tool selection, and step-by-step problem-solving across complex technical tasks.

Machine Learning Engineer at micro1, May 2025 - Aug 2025 · 4 mos, Remote, Contributed high-quality data annotation, evaluation, and model validation workflows for frontier AI systems. Performed data preprocessing and supported machine learning model development using TensorFlow and Scikit-learn.

Member at Department of CSA, IISc, Jan 2025 - Jan 2026 · 1 yr 1 mo, Bengaluru, Karnataka, India, Presented complex logic puzzles and computational theory to over 30,000 visitors, making challenging concepts accessible and increasing departmental visibility.

Recreational Coordinator at Rhapsody, IISc, Aug 2023 - Mar 2026 · 2 yrs 8 mos, Bengaluru, Karnataka, India, Hosted inter-college gaming tournaments, managing the full event lifecycle for over 200 registered teams. Administered a Discord community of 500+ members to streamline communication, registration, and technical support during the event.

Education

Bachelor of Technology - BTech, Mathematics and Computing, Indian Institute of Science (IISc)

High School, St. Ann's School

Public links

github: https://github.com/vaibhav3000

leetcode: https://leetcode.com/NINJA3000

vaibhav3000.github.io: https://vaibhav3000.github.io/portfolio/