Aman Singh Thakur — AI Researcher

I'm a MLE 2 at Amazon. At AWS Sagemaker AI, I was building agentic workflows for model customization — spanning model fine-tuning, evaluation, and deployment using tools like PyTorch, LangChain, and the Agents SDK. I specialize in building scalable ML systems, with experience across model training and serving infrastructure, ETL and data pipelines, real-time anomaly detection, and multi-agent orchestration using MCP and A2A protocols. I also work on fraud detection systems and ML infrastructure at scale. Previously, I worked as a Graduate Researcher at Meta, evaluating LLM alignment and judge reliability across models like Llama 2, Mistral, and GPT-4 for tasks including QA and ranking. I've also completed research internships at SLAC National Accelerator Laboratory, Stanford and Indian Institute of Technology (IIT), Kharagpur, where I developed ML-driven systems for online modeling, predictive analytics, and scientific simulation. Beyond research, I bring industry experience from Goldman Sachs and Morgan Stanley, where I built credit risk models, real-time anomaly detection systems, NLP pipelines, and large-scale data infrastructure powering some of the world's largest financial platforms. I hold an MS in Computer Science from UMass Amherst, with a focus on Distributed Systems, Deep Learning, and LLMs. I'm an active contributor to open-source, with Google Summer of Code projects at CERN-HSF and Lund University — building ML tools with PyTorch, Spark, and Autoencoders for scientific data compression and online physics modeling.

Stackforce AI infers this person is a Machine Learning Engineer with expertise in AI Research and Financial Technology.

Location: Seattle, Washington, United States

Experience: 4 yrs 9 mos