Role snapshot
Title
Tech Engineering Lead
Location
- Bengaluru, on-site
- Reports to
- Head of Engineering
Experience
6 to 8 years in software engineering, at least 2 as a
technical lead
Type
- Full-time, hands-on (most of your time in code and
- architecture)
- Open items before publishing: confirm work mode,
- compensation band and ESOPs.
About TalkingLands
- TalkingLands is a Bengaluru-based prop tech and AI startup,
- founded in 2021. We turn land records, cad astral maps, regulatory rules and
- satellite data into property intelligence that developers, lenders and buyers
- use to make land decisions. We already work with leading developers such as
- Prestige, and have more than 50,000 active users.
- The engineering problem is hard and genuine. We hold
- millions of land parcels as spatial polygons, map every property address to a
- location, and answer questions about containment (which village, zone or
- jurisdiction a parcel falls in), proximity (what lies within a regulatory
- buffer) and overlap (which restrictions touch a boundary). Answers must be
- traceable to a source record, because customers make purchases and lending
- decisions on them.
- Our platform is built on patented spatial intelligence.
- Technology and deep integration with Karnataka’s regulatory and cad astral systems.
The role
- This is a purely technical role in our engineering team. You
- will lead the architecture and code behind our AI and spatial platform, and
- make sure it stays fast and reliable as our data and users grow.
- You write production code every week, set the technical
- direction, and own the quality, performance and reliability of everything we
- ship. If you love going deep on AI systems, databases and scalable
- architecture, this role is for you.
Responsibilities
- AI systems
- •Design and build AI pipelines for document extraction
- using self-hosted small language models.
- •Build Agentic workflows and MCP-based tools on top of
- our spatial and property data.
- •Keep deterministic logic (registry lookups, spatial
- operations, regulatory calculations) separate from language-model steps, so
- every output can be cited and corrected.
- •Measure and improve model accuracy, latency and cost in
- production.
- Scalable platform
- •Design services, APIs and data models for spatial and
- document workloads, and write production code in them.
- •Keep spatial queries fast at scale: indexing, query
- plans, tile generation and caching.
- •Scale the platform for millions of parcels and a
- fast-growing user base without losing speed or reliability.
- •Review significant pull requests for correctness,
- clarity and performance.
- Quality and operations
- •Own test strategy, CI/CD pipelines and release quality.
- •Own production health: monitoring, on-call, incident
- response and postmortems.
- •Build security and audit requirements (access control,
- logging, data handling) into the architecture, not after it.
- •Track and pay down technical debt.
- Must have
- •6 to 8 years building and running production software,
- with at least 2 years as a technical lead.
- •Strong backend engineering in TypeScript (Node.js) or
- Python, including API design and service boundaries.
- •Deep PostgreSQL experience: schema design, indexing,
- query plan tuning and migrations on tables with millions of rows.
- •Hands-on experience building with language models in
- production, such as document extraction, retrieval or agents.
- •Has designed and scaled systems that ran in production
- for years, and can explain the trade-offs made.
- •Working knowledge of containers, Kubernetes and CI/CD,
- enough to debug a failed deploy yourself.
- •Bachelor’s degree in engineering or computer science,
- or equivalent experience.
- Good to have
- •PostGIS and spatial operations: point-in-polygon,
- buffers, intersections, coordinate reference systems.
- •Vector tile pipelines (Tippecanoe, PMTiles) and web map
- rendering with MapLibre GL JS.
- •Agentic workflows, tool use and MCP.
- •Self-hosting, serving or fine-tuning small language
- models.
- •Exposure to land records, cad astral data or real estate
- due diligence in India.
- •Experience working under SOC 2 or ISO 27001 controls.
- •GitOps deployment with ArgoCD.
- Tech stack
Layer
What we use
Data
- PostgreSQL with PostGIS as the system of record
- Map delivery
- Static PMTiles built with ogr2ogr and Tippecanoe, served
- over CDN
Frontend
TypeScript, MapLibre GL JS
Services
TypeScript and Python
AI
Self-hosted small language models for document extraction,
MCP, Agentic workflows
Platform
- Kubernetes, ArgoCD, CI/CD pipelines
- Why join us?
- •Purely technical work: AI, spatial data and scalable
- systems, nothing else on your plate.
- •A hard, unsolved problem: making India’s land data
- reliable enough to buy and lend on.
- •Real users at real scale: leading developers like
- Prestige, and more than 50,000 active users.
- •Ownership of the architecture from day one, and
- technical challenges that grow as we grow.
- If you love hard technical problems in AI and scalable
- systems, we would love to hear from you. Send your resume and a short note on a
- system you designed and what you would do differently today. Application email
- to be added.
Job description provided by Talking Lands.
View the original posting