Open to opportunities

prem kumar

@premkumar12

Message

Director of Engineering building production AI search, RAG, and agentic automation platforms.

India

Message

What I'm looking for

I’m looking for a role where I can own architecture and delivery end-to-end, lead cross-functional teams, and build production AI search/RAG and agentic automation with strong reliability, observability, and cloud cost focus.

I’m a Director of Engineering with 10+ years of progressive experience, moving from software engineer to executive ownership. I lead end-to-end architecture, product roadmap delivery, system reliability, and cloud cost optimization—always hands-on with AI/ML and production engineering.

At Instant Systems (Oct 2019–Present), I lead a 12–15 member cross-functional team across backend, frontend, AI/ML, DevOps, and QA. I oversee technical direction, delivery execution, observability, SLA/SLO ownership, and engineer coaching—driving operational excellence and innovation.

I build scalable AI search and GenAI systems, including production hybrid search (BM25/LTR + vector retrieval), rerankers, and context engineering. I also lead RAG and evaluation (LLM-as-Judge, Langfuse, TruLens, RAGAS) and develop agentic workflows using tools and agent frameworks like LangChain and Gemini/OpenAI-style models.

My platform work spans document intelligence and agentic acquisition—processing millions of documents and integrating thousands of external sources—while transforming unstructured inputs (PDFs/scans/OCR/Word/Excel) into structured entities. Earlier, as an AI Vision Engineer, I owned the full vision AI lifecycle from dataset curation to production-candidate model evaluation, including computer vision systems for classification, detection, and segmentation.

Experience

Work history, roles, and key accomplishments

Current

Director Engineering RAG Evaluation

Current

Instant Systems

Oct 2019 - Present (6 years 9 months)

Led the end-to-end engineering organization for a search and ranking platform, overseeing a 12–15 member cross-functional team across backend, frontend, AI/ML, MLOps, DevOps, and QA. Architected and scaled production hybrid search (BM25 + vector retrieval with reranking) and LLM/RAG-based document intelligence pipelines, including evaluation and reliability ownership.

Full Stack & ML Engineer

Vianai Systems

Oct 2022 - May 2024 (1 year 7 months)

Delivered core ML engineering for enterprise AI products, building model-training, evaluation, and productionization pipelines and integrating ML workflows into customer-facing applications. Shipped full-stack capabilities and production systems using cloud, containers, and CI/CD tooling.

AI Vision Engineer

Clove Dental

Jul 2020 - Oct 2021 (1 year 3 months)

Owned the Vision AI lifecycle, curating and annotating datasets, designing augmentation and transfer-learning workflows, and training CNNs for dental-image classification, object detection, and segmentation. Built evaluation workflows (precision, recall, mAP) and iterated models from prototype to production-candidate quality.

Computer Vision Data Annotation Transfer Learning CNN Object Detection Segmentation Python OpenCV TensorFlow PyTorch YOLO Mask RCNNs Precision Recall Maps Google Colab Matplotlib

Software Engineer

Instant Systems

Oct 2015 - May 2019 (3 years 7 months)

Built full-stack business-document search modules and web-acquisition frameworks, progressing from delivery work into architecture ownership. Engineered document-intelligence and NLP pipelines and delivered search-related backend services and REST APIs with monitoring and deployment ownership.

Legal Case Management

National Informatics Centre (NIC)

Collaborated on the development and enhancement of a legal case management system by optimizing data workflows and supporting digitization of legal documents. Produced reports and insights to streamline case handling and improve tracking efficiency.