Miju Labs

← All buyers · Google DeepMind

H

Hansa Srinivasan

Engineering lead, Nano Banana (Gemini Image), Google DeepMind · Google DeepMind

low confidenceTier A · Talk to firstFrontier labsimage

Engineering lead on Nano Banana, and the person who has stated the measurement problem most plainly: faces and aesthetic quality are very hard to quantify, so benchmarks are inadequate and the team falls back on attention to detail and human judgement. She is the technical counterpart to Nicole Brichtova and the person who would have to integrate any purchased evaluation data into the training loop.

Background

Srinivasan is an engineer at Google DeepMind who led engineering on the Gemini Image / Nano Banana line. Her public appearance is the Sequoia podcast with Brichtova on how the model achieved character consistency, where she is candid that the qualities users actually care about — likeness, aesthetic quality — resist quantification, and that the team's method was obsessive focus on specific problems such as text rendering rather than a general metric.

That is a rare admission from an engineering lead and a useful one: it means the internal evaluation stack for the most successful image model of 2025-26 rests on human judgement by design, not by omission.

What they run now

Engineering for the Gemini Image model family: training and evaluation infrastructure, the specific capability pushes (character consistency, text rendering) and the human-evaluation processes the team trusts over benchmarks.

What they buy

Inference, anchored in her own statement of the problem. What an engineering lead in her position needs is graded evaluation sets that behave like benchmarks but encode human judgement: stable, versioned, with per-item rationale and known inter-rater agreement, so a model change can be scored without convening a room of people every time. That is a different product from raw preference pairs and a more defensible one. Also plausible: adversarial or diagnostic sets for specific failure modes (text rendering at small sizes, identity drift across edits).

Career

  1. presentSoftware engineer → engineering lead, Gemini Image (Nano Banana), Google / Google DeepMind

On the record

Limits of benchmarks

Faces and aesthetic quality are things that are very hard to quantify, so benchmarks fall short. source

What drives quality

Progress came from attention to detail and the quality of everything you do with the model, rather than a single metric. source

Interviews & talks

Why it matters here

Medium-high. She is not the buyer, but she is the person who decides whether purchased data is usable. A pitch that survives her — versioned, reproducible, with agreement statistics — is a pitch that survives the whole Google DeepMind image org. She is also the natural second contact after Brichtova, and naming her specific quote is a credible way to show Julian has done the homework.

How to reach

Technical and narrow. Speculation: offer to build one diagnostic eval set for a failure mode she has named — identity preservation across multi-turn edits, say — with human grades, rationale and agreement statistics, and let it be tested against their internal results. Approach via LinkedIn or alongside Brichtova rather than cold.

What we could not establish
  • Exact title and tenure not evidenced beyond 'Engineering Lead' in the podcast billing.
  • No X handle or personal site found.
  • Education and career history unknown.
  • Date of the Sequoia episode not captured.

Why they are on the map

Nano Banana engineering lead on Sequoia 'Training Data' podcast; credits meticulous data curation; Imagen 3 core contributor source

Needs the data for a model; writes the spec.

How Google DeepMind buys data

Imagen 3 report (2024-08) collected 366,569 ratings from 3,225 raters in 71 countries via an unnamed 'external platform' under GDM data-enrichment policy (living wage). Veo 3.1 page cites human side-by-sides on MovieGenBench (1,003 prompts), VBench I2V and internal sets. Nano Banana leads (2025) describe large-scale human evals plus trusted testers and artist reviews; Gemini Data team job post (2026) covers data curation/generation/evaluation.

Aug 2026: Kavukcuoglu runs GDM incl. Gemini as SVP; Hassabis chair. Gemini has Data, Post-Training, Safety Post-Training and Model Behavior teams (job posts); Gemini app sits inside GDM under Josh Woodward. Gen-media (Imagen/Nano Banana/Veo/Flow) has product lead Brichtova and modality research leads (Wang, Erhan).

Sources

  1. How Google's Nano Banana Achieved Breakthrough Character Consistency
  2. Hansa Srinivasan, LinkedIn