A founding-team role to build HiLo's own large language model (LLM) and retrieval-augmented generation (RAG) stack.

Posting details

Company: Hilo AI Labs (HiLo) — hilo.finance

Location: Hyderabad — on-site

Employment type: Full time

Seniority: Senior / Staff, by experience

Experience: 5+ years in ML or software engineering, including 2+ years building LLM applications

Salary: ₹15–40 LPA, plus equity and a GPU and tooling budget

About HiLo

HiLo is an AI-powered chat platform where retail traders, investors and verified mentors talk markets in one place. The best trading edge is a connected, informed community, backed by AI that reads the tape with you.

- HiLo app: Direct messages, group channels, live audio rooms, premium paid channels, ticker cards, voice memos and trade journaling in the HiLo app (iOS, Android, Mac and Windows).
- Mentor marketplace: Verified mentors with brokerage-linked positions, transparency scores, trade alerts and paid 1-to-1 mentorship.

- Podcasts: Creator-led market shows inside the app.
- - Market data and charting: Stocks, exchange-traded funds (ETFs), options chains and multi-leg strategies.
- - HiLo Intelligence: Proprietary AI that turns filings, flow and price action into plain-English insights, catalyst and risk callouts, daily audio market briefings, and AI summaries of rooms and podcasts.
- Learn more at hilo.finance.

The role We are hiring a founding-team AI / ML Engineer to build HiLo's own LLM and retrieval (RAG) stack, so HiLo Intelligence runs on models and data we control. Today our AI features summarise filings, rooms and podcasts and flag catalysts and risks.

Next, we are building a finance-tuned LLM and a RAG layer over market data, SEC filings, options flow, transcripts and HiLo's own community conversations — answers that are grounded, cited, fast and cheap enough to run on every chat, room and ticker in the app. You will own this from first prototype to production, working directly with the founders, mobile and backend engineers.

What you'll do

Build HiLo's LLM

- Choose and fine-tune open-weight base models (supervised fine-tuning, LoRA/QLoRA, DPO) for finance: filings, earnings calls, options, trader slang and tickers.




- Build the data pipeline: collect, clean, dedupe, label and version training data, including synthetic data and expert feedback from HiLo mentors.

- Distil and quantise models to serve at low latency and low cost per query.

Build the RAG platform
- Design ingestion and chunking for SEC filings, news, earnings transcripts, market and options data, podcasts and chat history.
- Run hybrid retrieval (vector + keyword), re-ranking and freshness handling for fast-moving market data.

- Return grounded answers with citations to the filing, room or message they came from.
- - Add agent and tool calling: live quotes, options chains, charts and portfolio context.

Ship and run it in production
- Serve models with vLLM, TGI or similar on GPU infrastructure; manage cost, caching and autoscaling.
- Power features end to end: room and podcast summaries, daily audio briefings, catalyst and risk callouts, and an in-chat AI market assistant.
- Build evaluation: golden datasets, hallucination and citation checks, finance accuracy tests, A/B tests and online metrics.
- Add guardrails: no personalised financial advice, compliance filters, personally identifiable information (PII) protection and safe handling of user conversations.
- Make build-versus-buy calls on third-party model APIs versus our own models, and set the AI roadmap with the founders.

What you bring

- 5+ years in ML or software engineering, with 2+ years building LLM applications that real users rely on.
- Hands-on fine-tuning of open-weight LLMs (Llama, Mistral, Qwen) with PyTorch and Hugging Face, and a feel for when fine-tuning beats prompting or RAG — and when it doesn't.
- At least one production RAG system shipped: embeddings, chunking, vector databases (pgvector, Pinecone, Weaviate, Qdrant or similar), hybrid search and re-ranking.





- Strong Python and solid backend skills: APIs, asynchronous services, queues and data pipelines.
- - Experience serving models on GPUs (vLLM, TGI, Triton or similar) on AWS, GCP or Azure, with Docker and Kubernetes.
- - A habit of measuring: you build evaluations before you call something better.
- - Clear writing, and comfort in a small, fast team where you own decisions end to end.

Nice to have
- Finance or fintech experience: SEC filings (10-K, 10-Q, 8-K), earnings transcripts, market-data feeds and options analytics.

- Pre-training or continued pre-training at scale, or distributed training (DeepSpeed, FSDP).
- - Speech and audio: speech-to-text for live rooms and podcasts, text-to-speech for audio briefings.
- - Real-time and streaming systems (WebSockets, Kafka) and chat products.
- - Agent frameworks and tool calling (LangGraph, LlamaIndex, or your own).
- - Finance compliance and responsible AI (SEC/FINRA communications rules, avoiding personalised advice).
- - Open-source contributions, papers or public projects in LLMs or retrieval.

What success looks like
- First 30 days: Audit current AI features, data sources and costs, and set up an evaluation suite with baseline scores.

- By 90 days: RAG v1 in production — in-chat answers on filings and tickers, with citations.
- - By 6 months: First HiLo fine-tuned model live on at least one feature, beating the API baseline on quality or cost.
- - By 12 months: HiLo's own LLM and RAG powering most AI features, with clear quality, latency and cost metrics.

Why join HiLo
- Build a company's core AI from scratch, not maintain someone else's.
- - Real users and real data from day one: live rooms, mentors, podcasts and market feeds.
- - Founding-team ownership, with direct influence on product and AI strategy.
- - Competitive salary and equity, plus a GPU and tooling budget.

How to apply Apply via LinkedIn, or send your resume and a short note on an LLM or RAG system you built to [email protected] — what it did, how you measured it, and what you would change. GitHub, papers or demo links are welcome.

HiLo is an equal opportunity employer. We welcome applicants of every background.

📌 AI / ML Engineer — LLM & RAG (Hyderabad)
🏢 HiLo Platform
📍 Hyderabad