Headline

Lead Architect | AgenticAI, Evals | Python, FastAPI, Azure, AWS | OpenAI Agent SDK | Claude Certified Architect-Professional

About

A Forward Deployed Engineer (FDE), someone who works inside the client's environment, builds against real data and legacy systems, and owns the outcome. I stand out as a rare blend of Engineering, Systems integration, Cloud, LLM tooling and the ability to commuicate with client, define the acceptance criteria and deliver the desired outcomes. The one who can deliver the AI Outcomes and ensure ROI. I build business solutions that provide high value using AI. I solve problems that need to derive meaning out of unstructured data, right from sourcing data, to refining and training models and eventually consuming these implementations at scale. I delivered projects and insights to the clients which realized value in the ranges of $5mn to $50mn+ in short duration for business operations. I have lead high impact teams that convert vison into value through technology for clients. Managed team size of 16 spread across two continents, combined with vendor augmentation in delivering tech solutions. My tech capabilities include - Python, Langchain, LangGraph, Dotnet core, Javascript, Azure, Kubernetes, Serverless, AI copilot, Qdrantdb, SQL DB, DB2, Replicate, Comfyui, stable diffusion, Runpod, GPU accelerated sol. Libraries, frameworks - NLTK, Scikit-learn, pandas, Langchain, pytorch, tensorflow, OpenAI, Huggingface, Llama, Qwen models, sentence transformers. Domains - LifeSciences, Pharmaceutical Fin ops, Banking, and Provider healthcare.

Location

Hyderabad

AI stack

hugging face

langfuse

mcp

openai

DSPy

OpenAI Agents SDK

Work experience

Computational Science Sr Manager at Accenture, Feb 2026 - Present · 10 mos, Hyderabad, Telangana, India, Lead Architect - North America clients: Own enterprise AgenticAI architecture for Fortune 500 NA accounts; design RAG platforms, agentic systems, and LLM orchestration spanning data pipelines, model serving, observability, mutiple datastore and governance on Snowflake, Azure and AWS. Leadership- Lead and mentor a cross-functional team of 25+ AI/ML, NLP, Anaysts and MLOps engineers across multiple client engagements; set data management standards, review architectures, and grow senior IC talent. Enterprise agentic platform: Architected a hierarchical supervisor–worker agent framework for a client cognitive-AI platform, implementing Document drafting tools, memory management, guardrails, streaming, and human-in-the-loop controls; resolved concerns around LLM Vs Procedural tool builds to resolve bottlenecks and precision. Edge & efficient inference- Delivered quantized and fine-tuned LLMs (Llama 2/3, Mistral, Mixtral, Falcon, Zephyr, Cohere) for edge and latency-sensitive use cases using PEFT/LoRA/QLoRA; benchmarked with ROUGE, BLEU, DeepEval, and cosine similarity. Pre-sales & thought leadership- Represent Accenture in NA client architecture reviews and solution pitches; Provided recommendations on decisions pertaining to Microsoft Copilot to Microsoft Foundry of agents.

AI Expert at Ciklum, Sep 2022 - Feb 2026 · 3 yrs 6 mos, I design and build AI solutions that utilize LLMs and make our clients work easier and thereby turn our client's business more profitable. We use OpenAI at large as an LLM provider and also opensource LLMs like Llama, and various new benchmark leader LLMs. Our solutions are highly flexible and can be deployed on GPU powered clouds at ease like Azure, AWS, Replicate and Runpods. My recent works include (a) Deliveries on Azure landscape, where I enabled our clients with usage of Azure AI services, more specifically AI search, Azure Document understanding and Language. Effectively leading to 7 functional teams being able to build and maintain their own internal doc comprehension apps. (b) Custom document comprehension apps that integrate into custom data sources like Confluence, Relational Databases and Webscrapers. Crunching down a talent fulfilment cut down their efforts by 17%, measured by rate of onboarding speed. (c) Leading cross functional teams to achieve a problem solution fit, and provide a learning experience for the individuals, plus grow at the same time. I'm actively involved in upgrading frameworks for our clients that can speed up their frictionless and secure AI adoption journey. We strategically align with Client's larger vision and recommend roadmaps that emphasize on initiatives, execution and measuring methods. On the other spectrum I lead sales and deliveries across Ciklum for Conversational AI platform vendor partners like KoreAI, Laiye etc. Tools and services used- Azure AKS, AWS EC2, Chrome extensions, Qdrant, Langchain, Azure AI, Copilot,JS, Python. I also enjoy hosting Hackathons and conducting seminars internally and externally on topics around Solutioning of AI apps.

Senior Architect(Conversational AI, Analytics, ML Ops) at HCL Technologies, Jul 2021 - Sep 2022 · 1 yr 3 mos, Hyderabad, Telangana, India, - Provided consultation to architect business-vertical based segregation of Microservices. - Recommend architectural improvements, design solutions and integration solutions. - Leveraged Azure PaaS and SaaS, to improve Contact center(CX)/CCAI/Conversational AI engagements. - Evaluate network strategy and standards WRT to business and coordinate the implementation. - Provided solutions for new functional designs and optimized products for performance improvement. - Held design workshops with systems partners and stakeholders to realize the value in analytics through structured goal settings. - Lead a team of designers and development leads to implement the architecture. - Handled creative projects from concept to completion while managing vendor teams. Everyone must understand others thoughts! Hence I was able to bring balance to chaotic architecture diagrams into more fluidic versions through Miro.

Tech Arch Associate Manager at Accenture, Feb 2021 - Jun 2021 · 5 mos, Hyderabad, Telangana, India, - Lead the proceedings and operations of the Accenture AI Studio at Hyderabad - Provided the "first-step" experience into AI for C suit executives of leading orgs - Lead pre-sales initiatives in "AI in Ed-tech" and "AI in Surveillance" - As a Microsoft Cloud Architect, lead and implemented AI solutions using Azure AutoML and Conversational AI.

Development lead at Accenture, Dec 2016 - Feb 2021 · 4 yrs 3 mos, Greater Hyderabad Area, -Architected cutting edge Botframeworks for large enterprise clients in the domains Energy and Healthcare. - Helped scale the solution of the customers across his customer base. -Worked on several Enterprise integration for Microsoft Teams and Microsoft cognitive services. -Lead a highly motivated and diverse team, consisting of developers and BAs.

Software Engineering Sr. Analyst at Accenture, Sep 2014 - Dec 2016 · 2 yrs 4 mos, Greater Hyderabad Area, Client: US based Healthcare giant -Worked on several RFPs arising out of Data and AI propositions for Healthcare customers. -Delivered software solutions through "Distributed" Agile Methodology -Selected, on-boarded and mentored relevant technical expertise to deliver high quality for customers

Senior Software Engineer at HSBC Global Technology India, May 2013 - Sep 2014 · 1 yr 5 mos, Built and enhanced various features of HSBC Connect, which is a sophisticated application of the bank that enables it's enterprise customers to support voluminous payments of varied amounts of value, currencies and speed of transfer.

Software Engineer at HSBC Global Technology India, Jul 2009 - May 2013 · 3 yrs 11 mos, Pune District, Maharashtra, India, Built out several payment mapping applications that enable customers use their in-house ERP formats to accomplish payments through the bank. Predominantly worked on enabling these payments for customers in Europe.

Education

Masters in Data Science C30, Data Science, Liverpool John Moores University

EXECUTIVE POST GRADUATE PROGRAMME, DATA SCIENCE, International Institute of Information Technology Bangalore

Public links

github: https://github.com/saikumaru

saikumar.uk: https://saikumar.uk/