Principal AI Engineer

TekWissen LLC; Bengaluru; On-site; INR 1,653,000 - 3,031,000; Part time

Job summary

TekWissen Group is seeking a Consultant, IT Architecture in Bangalore to build AI-native security workflows on onsite delivery. You will work from approved designs, replace manual steps, and coordinate with functional owners to define data sources and edge cases.

The role emphasizes production-grade engineering, observability, and cost-aware deployment across engineering teams.

Qualifications

• 12+ years in software engineering with production ownership.
• Production-grade AI-native security workflow development.
• Strong programming in Python and TS/JS with testing and review practices.
• Experience with REST/GraphQL APIs, OAuth, and observability.

Responsibilities

• Build AI-native security workflows from approved designs.
• Collaborate with owners to define data sources and edge cases.
• Re-engineer processes for automation and consolidation.
• Ensure observability, cost controls, and reliable production delivery.

Skills

Python

TypeScript/JavaScript

Agentic systems

REST & GraphQL

Observability & delivery

Education

Bachelor's or Master's in CS

Tools

MCP servers

GitHub Copilot

Claude Code

LangGraph

Job description

Overview

TekWissen Group is a workforce management provider operating throughout India and several other countries worldwide. The client below is a leading technology company offering a range of IT solutions tbusinesses and organizations, enabling them ttransform their digital futures

Position: Consultant, IT Architecture

Location: Bangalore

Job Type: Contract

Work Type: Onsite

Job Description

• We are seeking a Senior Principal Engineer to build the AI-native capabilities that replace how security work is performed today.
• Security processes today run across multiple functional teams, frequently duplicated with different tooling and different definitions of done. This role re-engineers and implements the rationalized version of those processes as AI-native workflows, working from an approved solution design and directly with the teams whose work is being agentified.
• The expectation is production-grade engineering. Systems are delivered tested, instrumented so that the agent operations team can monitor them, and documented sufficiently for another engineer to extend them. Defects found in production are resolved by this role at root cause.

Objectives of This Role

• Cross-functional security workflows delivered as working AI-native implementations that replace the manual and fragmented processes they came from
• One implementation serving every function that shares a process, preserving in the build the consolidation the design sets out
• Delivery on committed timelines, with the reliability, observability, and cost profile to run in production
• Reusable components and patterns, so each workflow after the first starts from existing code

Responsibilities

• Build the AI-native implementation of security workflows from approved solution designs, including agent loops where an agent is the right pattern, tool and service implementations, sub-agent orchestration, state and checkpoint handling, and the deterministic components that own gating and verdict logic
• Work directly with the functional owners and practitioners whose process is being agentified to establish the operational detail a solution design does not carry: real data sources, exception paths, edge cases, and current decision logic
• Re-engineer the process as part of implementing it, rather than reproducing it step for step: remove steps that exist only because the work was performed by hand, collapse handoffs and queues, and restructure sequence and timing where agentic execution changes what is possible
• Implement the consolidated workflow so that every function sharing the process is served by one build, rather than forking the implementation per team
• Implement context strategy in code: retrieval and search, memory, grounding against enterprise data sources, prompt assembly, schema-constrained output, and context budgeting under real payload sizes
• Build tool and integration surfaces against the systems each function already works in, including MCP servers, API clients, and event consumers, with scoped credentials, rate limiting, retries, backpressure, and idempotency for actions that cannot safely repeat
• Implement automated tests, validate agent behavior against the acceptance criteria defined for the workflow, and hold to code review, typed interfaces, dependency hygiene, and documentation sufficient for another engineer to extend the system
• Instrument systems for operations, providing execution traces spanning agent steps and tool calls, token and cost accounting, and latency measurement that the agent operations team can monitor against
• Engineer for cost and latency through prompt and response caching, model selection and routing with fallback, and context compaction, with measurement to confirm the effect
• Harden systems against adversarial input, including prompt injection through retrieved or user-supplied content, tool abuse, and unintended data movement through agent surfaces
• Resolve defects escalated from agent operations, including non-deterministic behavior and degradation that surfaces only under production load, fixing at root cause with regression coverage that prevents recurrence
• Validate the delivered workflow with the functional owners before release, deliver within committed windows, and raise design gaps to the solution owner rather than resolving them unilaterally in code

Required Skills and Qualifications

• 12+ years in software engineering with direct ownership of systems in production
• Agentic systems you personally built and shipped into production, not prototypes, notebooks, or proofs of concept
• Strong programming proficiency in Python and at least one of TypeScript/JavaScript, Go, or an equivalent language, with demonstrable code quality practice covering testing, review, refactoring, and maintainability
• Hands-on implementation experience across:
  • Agent loops and tool calling: loop control, sub-agent delegation, state management and checkpointing, and idempotent handling of consequential actions
  • Context engineering: retrieval and search, memory, grounding in enterprise data, prompt assembly, and structured or schema-constrained output
  • Testing and validation: automated tests for behavior that varies between runs, regression suites, and validation against defined acceptance criteria
  • Cost and latency engineering: token accounting, prompt and response caching, model routing with fallback, and context compaction
• Process re-engineering experience: analyzing an existing workflow and rebuilding it around what automation makes possible, rather than encoding current practice into software
• Experience replacing an existing manual or fragmented business process with software, working directly with the people who perform that process today
• Integration engineering: REST and GraphQL APIs, event-driven consumers, OAuth and scoped credential handling, schema and contract versioning, and resilient behavior under partial failure
• Observability and delivery engineering: structured logging, distributed tracing, and metrics for systems whose behavior varies between runs, plus CI/CD pipelines, automated quality gates, and containerized deployment
• Ability to work independently from written specification, deliver against committed timelines, and communicate progress and blockers without prompting
• Bachelors or masters degree in computer science or a related field, or equivalent practical experience

Preferred Skills and Qualifications

• Experience applying AI within a security domain such as application security, DevSecOps, vulnerability management, code analysis, or software supply-chain security
• Experience building MCP servers or agent tooling, or extending AI coding assistants such as Claude Code, GitHub Copilot, or Cursor
• Working knowledge of agentic frameworks and orchestration libraries, and judgment about when to use them
• Experience with LLMOps in production, covering versioning, staged rollout, and regression detection
• Familiarity with the OWASP Top 10 for LLM Applications and with secure development practice
• Experience ramping quickly onto an unfamiliar codebase and becoming productive in it

Mandatory Skills

• Technical Skills: Python, LangGraph, LangFuse, LLM must haves + Typescript/JavaScript. Hands on Production based Implementation experience is mandate.
• Bachelor’s or master’s degree in computer science or a related field, or equivalent practical experience

Experience

• Total Exp:12+
• Rel Exp: 12+

TekWissen® Group is an equal opportunity employer supporting workforce diversity.