Jaswanth Kunisetty profile photo 1 on Tal

Jaswanth Kunisetty

AI Engineer II @ Stealth Startup | Ex-MLE @ Thoughtworks | AI & SDE

Bengaluru

3+ years of work experience

Stealth Startup
May 20264 mosPresent

Remote

Thoughtworks
Jan 20264 mosApr 2026

• Led a repo-wide MLOps v2 migration, enabling unified deployment of Databricks pipelines, GenAI services, and containerized micro-services in a single codebase; reduced cross-repo sync/maintenance effort by ~60%. • Automated post-pipeline validation and stakeholder reporting using AWS Lambda + EventBridge, replacing manual verification + email workflows; improved decision turnaround from hours to minutes. • Designed and deployed a Flask + Kubernetes review workflow app integrating with SQS queues and API events. Adopted by ~40+ stakeholders across 14 markets, eliminating manual SQS-based review cycles and cutting approval time by ~80%.
Thoughtworks
Jan 20251 yr 1 moJan 2026

Bengaluru

• Optimized Spark-based Databricks ML pipelines on AWS, reducing runtime by ~40% and accelerating marketing analytics delivery across 14 live countries, directly improving stakeholder reporting SLAs. • Supporting large-scale ML platform modernization for a global pharmaceutical client, spanning pipeline optimization, MLOps v2 migration, automation, and GenAI-driven document intelligence. • Contributed to a GenAI + graph-augmented retrieval PoC for intelligent search over 5K+ GxP documents, demonstrating cross-reference graph building + RAG-based semantic retrieval for domain experts. • Built strong collaboration channels with Data Science, DPH, and product teams, proactively identifying blockers during Germany/Italy go-lives and the US revamp, preventing delays and migration risk across markets.
HyperWorks Imaging
Aug 20244 moDec 2024

Bengaluru, Karnataka, India

• Engineered a multi-model OCR + CV + DL pipeline for lithology documents (TIF/PDF >200MB), achieving <5% extraction error tolerance across numeric and graphical data used in geological analysis. • Segmented depth curves, tabular logs, geological annotations and structural graphs using custom UNet/YOLO-based models, followed by OCR correction with ML-driven post-processing. • Built a multi-stage inference system combining graph segmentation + numeric extraction + text normalization, sustaining high fidelity across 4K+ complex documents. • Automated full document digitization workflow, eliminating ~90% manual inspection effort and enabling downstream analytics for domain experts. • Contributed to a RAG-based GenAI retrieval system using FAISS + LLMs to surface insights from digitized lithology datasets with ~84% faster information lookup vs manual search.
BNP Paribas CIB
Jan 20246 moJul 2024

Chennai, Tamil Nadu, India

• Developed a ServiceNow-integrated chatbot that automated ticket creation, reducing manual navigation efforts by >70%. • Automated scheduling & document-processing tasks using Python + Selenium, saving 5–7 hours/week of repetitive work. • Created a real-time JIRA analytics dashboard supporting release planning and delivery decisions with interactive data visualizations.
Innovation Incubator Advisory
Jul 20222 mosAug 2022

India

Developed and orchestrated a deep learning model for image denoising with an accuracy of 82.36%, using TensorFlow, Python, and Git, showcasing my expertise in AI and ML tools development.
Acceleron Labs
Oct 20217 moMay 2022

Bengaluru

Worked on speech recognition projects using NVIDIA NeMo libraries, contributing to the development of efficient speaker recognition models with up to 90% accuracy.

Jaswanth’s AI Stack

hugging face
hugging face
langchain
langchain
ollama
ollama
opencode
opencode
claudecode
claudecode
OpenRouter
OpenRouter

Education

Amrita Vishwa Vidyapeetham

Bachelor of Technology, Artificial Intelligence

Amrita Vishwa Vidyapeetham

20204 yrs2024

Join Jaswanth and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Jaswanth Kunisetty on Tal

Headline

AI Engineer II @ Stealth Startup | Ex-MLE @ Thoughtworks | AI & SDE

About

Driven by a passion for AI and software development, I love turning complex problems into impactful solutions. From digitizing centuries-old documents to creating intelligent systems, my journey has been fueled by curiosity and innovation. I'm eager to contribute my skills in AI and SDE roles to build technology that truly makes a difference.

Location

Bengaluru

AI stack

hugging face

langchain

ollama

opencode

claudecode

OpenRouter

Work experience

AI Engineer II at Stealth Startup, May 2026 - Present · 4 mos, Remote

Consultant - Machine Learning Engineer at Thoughtworks, Jan 2026 - Apr 2026 · 4 mos, • Led a repo-wide MLOps v2 migration, enabling unified deployment of Databricks pipelines, GenAI services, and containerized micro-services in a single codebase; reduced cross-repo sync/maintenance effort by ~60%. • Automated post-pipeline validation and stakeholder reporting using AWS Lambda + EventBridge, replacing manual verification + email workflows; improved decision turnaround from hours to minutes. • Designed and deployed a Flask + Kubernetes review workflow app integrating with SQS queues and API events. Adopted by ~40+ stakeholders across 14 markets, eliminating manual SQS-based review cycles and cutting approval time by ~80%.

Associate - Machine Learning Engineer at Thoughtworks, Jan 2025 - Jan 2026 · 1 yr 1 mo, Bengaluru, • Optimized Spark-based Databricks ML pipelines on AWS, reducing runtime by ~40% and accelerating marketing analytics delivery across 14 live countries, directly improving stakeholder reporting SLAs. • Supporting large-scale ML platform modernization for a global pharmaceutical client, spanning pipeline optimization, MLOps v2 migration, automation, and GenAI-driven document intelligence. • Contributed to a GenAI + graph-augmented retrieval PoC for intelligent search over 5K+ GxP documents, demonstrating cross-reference graph building + RAG-based semantic retrieval for domain experts. • Built strong collaboration channels with Data Science, DPH, and product teams, proactively identifying blockers during Germany/Italy go-lives and the US revamp, preventing delays and migration risk across markets.

Machine Learning Engineer at HyperWorks Imaging, Aug 2024 - Dec 2024 · 4 mo, Bengaluru, Karnataka, India, • Engineered a multi-model OCR + CV + DL pipeline for lithology documents (TIF/PDF >200MB), achieving <5% extraction error tolerance across numeric and graphical data used in geological analysis. • Segmented depth curves, tabular logs, geological annotations and structural graphs using custom UNet/YOLO-based models, followed by OCR correction with ML-driven post-processing. • Built a multi-stage inference system combining graph segmentation + numeric extraction + text normalization, sustaining high fidelity across 4K+ complex documents. • Automated full document digitization workflow, eliminating ~90% manual inspection effort and enabling downstream analytics for domain experts. • Contributed to a RAG-based GenAI retrieval system using FAISS + LLMs to surface insights from digitized lithology datasets with ~84% faster information lookup vs manual search.

Python Developer Intern at BNP Paribas CIB, Jan 2024 - Jul 2024 · 6 mo, Chennai, Tamil Nadu, India, • Developed a ServiceNow-integrated chatbot that automated ticket creation, reducing manual navigation efforts by >70%. • Automated scheduling & document-processing tasks using Python + Selenium, saving 5–7 hours/week of repetitive work. • Created a real-time JIRA analytics dashboard supporting release planning and delivery decisions with interactive data visualizations.

AI Intern at Innovation Incubator Advisory, Jul 2022 - Aug 2022 · 2 mos, India, Developed and orchestrated a deep learning model for image denoising with an accuracy of 82.36%, using TensorFlow, Python, and Git, showcasing my expertise in AI and ML tools development.

Python Developer Intern at Acceleron Labs, Oct 2021 - May 2022 · 7 mo, Bengaluru, Worked on speech recognition projects using NVIDIA NeMo libraries, contributing to the development of efficient speaker recognition models with up to 90% accuracy.

Education

Bachelor of Technology, Artificial Intelligence, Amrita Vishwa Vidyapeetham