Om Godhani

AI/ML Engineer

Actively looking · Member since 17 Sep 2026

Message

Location: Surat, India

Desired salary: Unspecified

Work preference: Remote Only

Experience level: Junior

About

Professional summary

I am an AI/ML Engineer focused on building and shipping LLM-driven features for B2B SaaS products. My work spans retrieval-augmented generation, agent orchestration, structured LLM output, prompt engineering, and evaluation systems.

I have built provider-agnostic LLM integrations across Claude, Gemini, Groq, and OpenAI-compatible APIs. I design reliable generation workflows with JSON-schema output, streaming, retry and backoff logic, context grounding, and quality checks.

I work with Python backend systems using Django, Django REST Framework, FastAPI, PostgreSQL, and REST APIs. I have implemented production features involving OAuth flows, encrypted credentials, webhooks, transactional integrity, and row-level locking.

I have experience creating evaluation harnesses that compare prompt variants against baselines and identify quality regressions using held-out test cases. I also build multi-stage content pipelines that use planning, drafting, reviewing, and revision steps to improve output quality and consistency.

My projects include role-based workflow APIs, an NLP financial-news sentiment classifier, web scraping and document parsing pipelines, and generative media workflows. I also have prior C++ backend experience working with trading Order Management System applications, file handling, sockets, and client-server architecture.

I am currently completing a B.Tech in Computer Science and Engineering with a specialization in Artificial Intelligence. I enjoy applying practical AI engineering principles to build reliable, testable, and scalable products.

Tech stack & tools

Working toolkit

Experience

Career history

AI Product Developer Vasundhara Infotech

Jun 2026 – Aug 2026

Owned the LLM content platform for Sensor Tower across five B2B SaaS products in the App Store and Google Play ecosystem. Designed a provider-agnostic LLM layer supporting Claude, Gemini, Groq, and OpenAI-compatible endpoints with structured JSON-schema output, streaming, and automatic retry and backoff handling.

Created an evaluation harness to compare prompt variants against baselines using held-out cases, binary pass/fail assertions, and aggregated scoring. Built a multi-stage self-critiquing generation pipeline using planning, drafting, editing, reviewing, and revision prompts grounded in brand knowledge and post history.

Developed a scraping and corpus-cleaning pipeline for 275 articles, extracting category-specific writing-style features and combining deterministic metrics with LLM-as-judge evaluation. Shipped publishing integrations for nine social platforms with OAuth 2.0, token refresh, media handling, and character-limit support.

Implemented Fernet-backed encryption at rest for OAuth tokens and API keys, including migration support and backend test coverage. Also produced SEO articles, generative brand media, and Postman API documentation for RefundSensor authentication, refund events, transactions, analytics, and Apple App Store Server Notification webhooks.

C++ Backend Developer Intern Wealth Affinity Softech

Jan 2026 – May 2026

Implemented file handling and cross-platform file access between Windows and Linux for a C++ trading application used in Order Management System operations.

Maintained Admin and Trader applications while debugging and navigating a large multi-file C++ codebase. Analysed Login Management, Order Management, and Strategy Mapping modules.

Worked in a client-server architecture using socket communication and real-time request handling.