Headline

Software Engineer @ InMobi | Forward-Deployed / Full-Stack AI Engineer: LLM products end to end, partner integrations, ad monetisation | Python · FastAPI · Next.js · ML

About

I'm a full-stack engineer at InMobi, and the work I like most is close to the people using the software: give me a fuzzy problem and a partner on the other end, and I'll get it from "we're not sure what this should be" to something shipped and running. Most of what I build sits where AI meets ad monetization, and I work end to end: the data model, the APIs, the LLM parts, and the interface. I've been most useful at the edges, integrating with outside partners and pushing things into production: - Wired up eight major ad networks and content feeds, with per-partner config so onboarding a new one is setup, not a rebuild. - Building an embeddable AI answer widget a partner can drop onto their own site with a single script tag. It grounds answers in the partner's own content (retrieval-augmented, with embeddings-based relevance scoring) and is origin-locked and isolated by design. - Built a bulk ad uploader that turns a partner's spreadsheet into thousands of live-ready ads, all created paused so nothing spends by accident. - Rebuilt an identity system from scratch (three-tier org hierarchy, role-based access, SSO, JWT) and built an LLM content-generation surface on Google Gemini and Anthropic Claude. On the production side: I cut auth latency up to ~8x with a Redis circuit breaker and caching, moved storage from AWS S3 to Google Cloud Storage, and did the hardening work that gets something launch-ready. Two of those I took from scratch to working software as part of a small team. Day to day: Python and FastAPI, PHP/Laravel, and TypeScript with Next.js and React, on Redis, Postgres/AlloyDB, Docker and GCP. I'm at my best close to a real user or partner, taking ambiguous problems to production. Always up for comparing notes with people building LLM products or working close to customers and partners.

Location

Bengaluru

AI stack

claude code

codex

cursor

hermes agent

opencode

Wispr Flow

Work experience

Software Development Engineer (SDE I) at InMobi, Jun 2025 - Present · 1 yr 3 mos, Bengaluru, Karnataka, India, • Helped build an AI monetization platform for publishers, advertisers and affiliates, from an MVP to a launch-ready product now moving from staging toward production. • Built or extended integrations across eight external ad networks and content feeds, with per-partner config so onboarding a new partner is setup, not a rebuild. • Built an LLM content- and keyword-generation surface on Google Gemini and Anthropic Claude (plus Imagen 4), with a prompt engine and a concept-discovery service. • Built an embeddable AI answer widget (prototype) partners add with one script tag; it grounds answers in a per-partner content manifest (retrieval-augmented) with embeddings-based relevance scoring, behind origin-allowlist auth and closed Shadow-DOM isolation. • Built a Meta Ads bulk uploader (Python CLI plus web app) that turns a partner's spreadsheet into up to 5,000 live-ready ads, all created paused so nothing spends by accident. • Rebuilt the identity system (IAM v3) from scratch: a three-tier org hierarchy with role-based access, SSO and JWT, across a Python/FastAPI backend and a Next.js/React frontend. • Wrote a Python ad-compliance service that validates campaigns against network policies: HMAC-signed requests, a review/approval workflow, and an LLM auto-approval layer. • Improved performance and launch-readiness: semantic image search over embeddings in the Image Hub, an up-to-~8x auth-latency cut (Redis circuit breaker plus caching), an AWS S3 to Google Cloud Storage migration, containerization (Docker, Jenkins, Helm), and security hardening.

Product Analyst at Glance, Jun 2024 - Aug 2024 · 3 mos, Bengaluru, Karnataka, India, • Built and maintained four Superset dashboards, including the USA-cluster Product Alert Metrics dashboard, tracking engagement across OEMs and regions for 120M+ users. • Ran regional deep-dives across 30+ engagement metrics and sent product managers bi-weekly insight reports; the work fed a 10% lift in retention and 15% in engagement. • Researched device demographics and smartphone usage (Xiaomi, Realme, Samsung) to guide content strategy for the US and Singapore clusters.

AI Project Intern at IBM, Jun 2024 - Jul 2024 · 2 mos, India, • Conducted advanced research and experimentation to enhance Natural Language Processing (NLP) models for real-world applications. • Explored and implemented state-of-the-art NLP techniques, gaining hands-on experience with transformers and language models. • Led the project on ’Plant Disease Detection and Prediction’ using machine learning algorithms to develop an AI-based solution for early detection and prevention in agriculture.

Summer Trainee at DRDO, Ministry of Defence, Govt. of India, Jun 2023 - Jul 2023 · 2 mos, Dehradun, Uttarakhand, India, During my summer break of B.Tech in Computer Science Engineering at the University of Petroleum and Energy Studies, I had the opportunity to undertake a significant research project titled "MAC Protocols in Tactical MANET" at IP Based Protocols Groups (IBPG), Defence Electronics Applications Laboratory (DEAL), Dehradun. The project spanned from 15th June 2023 to 26th July 2023. Under the guidance of Sh. Binod Pal, Scientist-'F' and Group Head- IBPG, and the mentorship of Sh. Vikas Yadav, Scientist-'E', I successfully completed the project with excellence. The experience provided me with valuable insights into MAC protocols and their applications in Tactical Mobile Ad-hoc Networks (MANETs). I am grateful for the knowledge and expertise gained during this training, and I am excited to leverage these skills in my future endeavours. I am looking forward to a successful career and a bright future.

Summer Research Intern at Indian Institute of Technology, Delhi, Jun 2023 - Jun 2023 · 1 mo, New Delhi, Delhi, India, I had the privilege of completing a research internship at IIT Delhi under the guidance of Dr. Anil Verma. The internship, spanning approximately 4 weeks from 1st June 2023 to 30th June 2023, focused on the intriguing subject of "Fault Detection and Diagnosis in PEMFC using Artificial Intelligence (AI)." During this enriching experience, I delved into the realm of AI-powered techniques for detecting and diagnosing faults in Proton Exchange Membrane Fuel Cells (PEMFC). Working closely with Dr. Anil Verma, I gained valuable insights into the latest advancements in this field and honed my research and analytical skills. I am grateful for the opportunity to contribute to cutting-edge research at IIT Delhi and for the guidance received from Dr. Anil Verma, whose expertise and mentorship were instrumental to my learning. This internship has fueled my passion for research and AI applications, and I look forward to leveraging this experience in future research endeavours.

Education

Bachelor of Technology - BTech, Computer Science, UPES

SSC, PCM, Grace Academy, Dehradun

Public links

github: https://github.com/SuryanshD