Sanat B Singh

AI Engineer & Tech Lead Hyderabad, India

World models, Agentic systems, and Applied Generative AI.

I build practical AI systems across product, research, and deployment, with a personal focus on controllable world models. At ETS GCC, the org behind TOEFL and GRE, I co-lead product development across interactive assessment, assessment simulation, agentic AI, and generative AI. Outside work, I actively research and experiment with causal, action-conditioned world models, and I am building Shortlist - an AI screening layer for early-stage investors. Previously at Deloitte USI, I worked as a Data Scientist and Machine Learning Engineer on LLM fine-tuning with DGX H100 clusters and edge computer vision pipelines on Jetson AGX Orin.

View work Resume Contact
LLM fine-tuning SLMs PEFT LoRA QLoRA DAPT RAG Multi-agent systems Agentic Systems World Models TensorRT TensorRT-LLM DeepStream PyTorch Hugging Face Nvidia NeMo DeepSpeed FastAPI Docker Kubernetes AWS Azure GCP Edge AI

Work

Current and recent systems work

Selected work summarized at a public-safe level, with enough detail to show the systems, models, deployment surfaces, and scale involved.

Sept 2024 - Present

AI Engineer, ETS GCC / Firstsource Solutions

Product research and development for ETS GCC Hyderabad, set up in collaboration with Firstsource Solutions. ETS is the org behind TOEFL and GRE.

  • Co-leading an end-to-end platform and framework/library for interactive assessment, spanning simulation development, skill practice, and deployment at scale.
  • Fine-tuning SLMs with PEFT for scoring and feedback workflows.
  • Designing multi-agent prompt generation, evaluation, and benchmarking systems for generative assessment use cases.
Generative AI Agentic AI SLM fine-tuning PEFT Multi-agent systems
July 2021 - Aug 2024

Data Scientist / MLE, Deloitte USI

Progression: Analyst, July 2021 - May 2023; promoted to Consultant, June 2023 - Aug 2024.

  • Fine-tuned LLaMA 2/3.1 and Mixtral 8x7B with LoRA, QLoRA, PEFT, DAPT, Hugging Face, Nvidia NeMo, and DeepSpeed on Nvidia DGX A100 and DGX H100 clusters.
  • Built finance assistant, RAG chatbot, and content generation systems with TensorRT-LLM, FastAPI, Docker, Kubernetes, AWS, and Azure deployment paths.
  • Shipped Edge AI and Computer Vision pipelines including ControlNet APIs and YOLOv8 traffic analysis on Nvidia Jetson AGX Orin with TensorRT and DeepStream.
  • Worked across VQA benchmarking, retail shelf management, MaskRCNN, AlexNet, ONNX export, and SAP-integrated model workflows.
LLM fine-tuning DGX H100 Jetson AGX Orin TensorRT DeepStream
July 2020 - July 2021

Research Intern, University of Houston Downtown

Worked with Dr. Hong Lin on EEG signal-based sleep stage classification and independent deep learning research for iRT and Ion Mobility prediction from peptide sequences.

2018 - 2021

Machine Learning Instructor, Konnexions KIIT

Taught and mentored around 120 undergraduate students in a semester-long Machine Learning course, including curriculum design and hands-on workshops.

Projects

Current personal builds

In development

Shortlist

An AI screening layer for early-stage investors. Research agents build investor thesis profiles, create personalized scoring rubrics, monitor inboxes, and run structured async founder conversations to surface signal from high-volume inbound.

The product combines deep research agents, persona modeling, inbox triage, and memo-style dashboards for early deal evaluation.

Agentic AI Research agents Evaluation rubrics Applied LLM systems

Timeline

Career timeline

Key milestones, with Deloitte progression split out for clarity.

  1. 2017 - 2021

    B.Tech, Computer Science and Engineering, KIIT Bhubaneswar

    Graduated with CGPA 8.99.

  2. 2018 - 2021

    Machine Learning Instructor, Konnexions KIIT

    Core team member, teaching and mentoring undergraduates in applied Machine Learning.

  3. Jan 2020 - May 2021

    ML Engineer and Co-founder, Zyik.ML

    Healthcare AI startup registered under MSME, Government of India, and backed by AWS Activate.

  4. July 2020 - July 2021

    Research Intern, University of Houston Downtown

    Deep learning research under Dr. Hong Lin, including EEG sleep stage classification.

  5. July 2021 - May 2023

    Data Scientist / MLE, Analyst, Deloitte USI

    Computer Vision, NLP, RAG, Conversational AI, and Edge AI delivery across enterprise projects.

  6. June 2023 - Aug 2024

    Promoted to Consultant, Deloitte USI

    Led more complex GenAI and deployment work across DGX A100/H100, TensorRT-LLM, Jetson AGX Orin, and cloud systems.

  7. Sept 2024 - Present

    AI Engineer, ETS GCC / FSL

    Interactive assessment, generative AI, SLM fine-tuning, and multi-agent evaluation systems.

Research & Publications

Publications and conference work

NARST Annual International Conference 2026

Using Multi-Student, Generative AI Teaching Simulations as Practice Spaces for Facilitating Science Discussions

Co-author on a stand-alone paper at the 2026 NARST Annual International Conference in Seattle, WA. Listed in the official conference program.

Book chapter and reviewer work

Zyik.ML - A CADx System; Guest Reviewer, Multimedia Systems

Book chapter in Project Innovation Contest 2021 Proceedings, and guest reviewer for Multimedia Systems, Springer Journal.

Recognition

Selected recognition

#11

VibeCon India 2026

Top 20 from 25,000 teams by Emergent, with Soumik Rakshit and Atanu Sarkar.

CBID

Johns Hopkins design challenge

Selected for the Johns Hopkins CBID Covid-19 Virtual Design Challenge as one of the few teams from India.

AWS

AWS Activate

Zyik.ML was selected for AWS Activate while building healthcare AI prototypes and CADx systems.

ETS GCC

Proactive AI ownership

Recognized at ETS GCC for proactive ownership and innovative work across AI product development.

Deloitte

Complex model delivery

Recognized for reliable turnaround on complex model work across Deloitte USI projects.

Contact

Open to thoughtful conversations.

Reach out about world models, agentic AI, generative systems, applied ML problems, or frontier AI engineering work.

Email LinkedIn GitHub Twitter/X Resume