Design, provision, and document a production-grade AWS micro-service platform for a Apache-powered ERP implementation—hitting our 90-day “go-live” target while embedding DevSecOps guard-rails the team can run without you. Key Responsibilities Cloud Architecture & IaC Author Terraform modules for VPC, EKS (Graviton), RDS (MariaDB Multi-AZ), MSK, ElastiCache, S3 lifecycle, API Gateway, WAF, Route 53. Implement node pools (App, Spot Analytics, Cache, GPU) with Karpenter autoscaling. CI/CD & GitOps Set up GitHub Actions pipelines (lint, unit tests, container scan, Terraform Plan). Deploy Argo CD for Helm-based application roll-outs (ERP, Bot, Superset, etc.). DevSecOps Controls Enforce OPA Gatekeeper policies, IAM IRSA, Secrets Manager, AWS WAF rules, ECR image scanning. Build CloudWatch/X-Ray dashboards; wire alerting to Slack/email. Automation & DR Define backup plans (RDS PITR, EBS, S3 Std-IA → Glacier). Document cross-Region fail-over run-book (Route 53 health-checks). Standard Operating Procedures Draft SOPs for patching, scaling, on-call, incident triage, budget monitoring. Knowledge Transfer (KT) Run 3× 2-hour remote workshops (infra deep-dive, CI/CD hand-over, DR drill). Produce “Day-2” wiki: diagrams (Mermaid), run-books, FAQ. Required Skill Set 8+ yrs designing AWS micro-service / Kubernetes architectures (ideally EKS on Graviton). Expert in Terraform , Helm , GitHub Actions , Argo CD . Hands-on with RDS MariaDB , Kafka (MSK) , Redis , SageMaker endpoints . Proven DevSecOps background: OPA, IAM least-privilege, vulnerability scanning. Comfortable translating infra diagrams into plain-language SOPs for non-cloud staff. Nice-to-have: prior ERP deployment experience; WhatsApp Business API integration; EPC or construction IT domain knowledge. How Success Is Measured Go-live readiness — Production cluster passes load, fail-over, and security tests by Day 75. Zero critical CVEs exposed in final Trivy scan. 99 % IaC coverage — manual console changes not permitted. Team self-sufficiency — internal staff can recreate the stack from scratch using docs + KT alone. Show more Show less
Test Strategy & Plan Map user stories (DPR, Mess QR, Leave, Incident, Material Req, HR) to acceptance criteria. Define the “shift-left” gates: PR smoke ➝ nightly regression ➝ pre-release performance ➝ post-deploy canary. Automated Functional Testing Build PyTest-BDD suites against ERPs REST UI & bot chat flows. Mock WhatsApp Cloud API via WireMock ; replay Meta payloads in CI. API Contract & Schema Tests Use Pact (Python) for Bot ⇄ Worker message schemas; enforce in GitHub Actions. Validate EventBridge / Kafka events against Confluent Schema Registry. Performance & Load k6 scripts for Bot API (5 k RPS) and Kafka throughput. Locust browser flows (100 concurrent users) against ERP Desk. Publish SLO dashboards in Grafana. Master Test Orchestrator Write a single Makefile / tox entry that: make ci → spins ephemeral EKS namespace ➝ seeds test data ➝ runs full suite ➝ tears down. Integrate with GitHub Actions reusable workflow. Chaos & DR Drills Inject RDS fail-over, MSK broker stop, and Redis node kill; assert that tests still pass or alert. SOP & Documentation Produce “QA Run-book” wiki: pipeline diagrams, how to add a test, how to debug failures. Checklist for each release: tag, run pipeline, sign-off matrix. Knowledge Transfer Conduct two 2-hour remote workshops: “Writing a new test in 15 minutes” and “Reading Grafana QA dashboards.” Deliverables & Timeline Zero critical regressions escape to production during 90-day window. < 15 min PR gate; nightly regression under 45 min. < 2 h / month maintenance on test scripts (measured by Jira log). All new features merge only when green master pipeline . Must-Have Skills Advanced Python test automation (PyTest, fixtures, mocks). Deep knowledge of CI pipelines on GitHub Actions. Load-testing with k6 / Locust and interpreting results. Experience testing micro-service event systems (Kafka, SQS, EventBridge). Clear technical writing—can turn pipelines into beginner-friendly SOPs. Show more Show less
What makes this role special Join a green-field Enterprise solutions project that spans cloud-infra, data pipelines, QA automation, BI dashboards and business process analysis. Spend your first year rotating through four pods, discovering where you shine, then lock into the stream you love (DevOps, Data Engineering, QA, BI, or Business Analysis). Work side-by-side with senior architects and PMs; demo every Friday; leave with production-grade experience most freshers wait years to gain. Rotation roadmap (three months each) DevOps Starter – write Terraform variables, tweak Helm values, add a GitHub Action that auto-lints PRs. Data Wrangler – build a NiFi flow (CSV → S3 Parquet), add an Airflow DAG, validate schemas with Great Expectations. QA Automation – write PyTest cases for the WhatsApp bot, create a k6 load script, plug Allure reports into CI. BI / Business Analysis – design a Superset dataset & dashboard, document KPIs, shadow the PM to craft a user story and UAT sheet. Day-to-day you will Pick tickets from your pod’s board and push clean pull-requests or dashboard changes. Pair with mentors, record lessons in the wiki, and improve run-books as you go. Demo your work (max 15 min) in our hybrid Friday huddle. Must-have spark Basic coding in Python or JavaScript and Git fundamentals (clone → branch → PR). Comfortable with SQL JOINs & GROUP BY and spreadsheets for quick analysis. Curious mindset, clear written English, happy to ask “why?” and own deadlines. Bonus points A hobby Docker or AWS free-tier project. A Telegram/WhatsApp bot or hackathon win you can show. Contributions to open-source or a college IoT demo. What success looks like Ship at least twelve merged PRs/dashboards in your first quarter. Automate one manual chore the seniors used to dread. By month twelve you can independently take a user story from definition → code or spec → test → demo. Growth path Junior ➜ Associate II ➜ Senior (lead a pod); pay and AWS certifications climb with you. How to apply Fork github.com/company/erpnext-starter, fix any “good-first-issue”, open a PR. Email your resume, PR link, and a 150-word story about the coolest thing you’ve built. Short-listed candidates get a 30-min Zoom chat (no riddles) and a 24-hr mini-task aligned to your preferred first rotation. We hire attitude over pedigree—show you learn fast, document clearly, and love building, and you’re in. Show more Show less
What you’ll architect & deliver (any module) Fleet & Equipment Management Vehicle / plant master, telematics feed hooks, fuel log, job-card workflow, preventive maintenance scheduler. Contractor & Sub-contractor Portal On-boarding wizard, compliance checklist, attendance capture, subcontract billing ledger, retention release triggers. Engineering / WBS Module Multi-level WBS tree, quantity-take-off import, DPR linkage, earned-value KPIs, design-change versioning. Asset & Tool Tracking QR / RFID scan, issuance-return cycle, depreciation, loss-damage workflow, cost-centre allocation. Safety & Incident Reporting Near-miss and LTI DocTypes, photo / video attachments, root-cause matrix, auto-escalation rules, OSHA dashboard. Vendor & Material Request Hub Catalogue master, three-way match, WhatsApp approval hooks, price-variance alerts, supplier scorecard. Mess & Welfare Services Daily QR menu, calorie tracker, guest authorisation, contractor cost-share ledger. HR-Time & Attendance Extension Geo-fence clock-in, biometric sync, multi-project timesheet roll-up, leave quota exceptions. Reporting & BI Glue Layer Structured datasets for Superset, drill-through links, row-level security tags, KPI dictionary YAML. Must-have expertise 5+ yrs deep Frappe stack (Python, Jinja, JS) – custom DocTypes, hooks, REST, background jobs. MariaDB query optimisation; Redis caching strategies. Git + CI pipelines (GitHub Actions or GitLab). Authoring migration / patch scripts for live, multi-company sites. Comfortable writing concise English docs & diagrams (Mermaid / Markdown). Nice-to-have EPC, Logistics, or Manufacturing domain exposure. WhatsApp Business API integrations. Experience with Kafka, EventBridge, or any streaming bus. Engagement details Timeline: 6–8 months (target go-live Month 3, hyper-care Month 4–5) Workstyle: 100 % remote, IST overlap 3 hrs daily. Pay: senior-market rate, monthly invoice; milestone bonuses on module sign-off. Stack: ERPNext v15 (Python 3.10), GitHub, Jira, Slack, VS Code, Docker. How to apply Send CV + GitHub/project link showing at least one custom ERPNext module. Briefly describe which module above excites you most and why you’re the right dev to ship it. Show more Show less
About the Role You’ll join a small, fast team turning cutting-edge AI research into shippable products across text, vision, and multimodal domains. One sprint you’ll be distilling an LLM for WhatsApp chat-ops; the next you’ll be converting CAD drawings to BOM stories, or training a computer-vision model that flags onsite safety risks. You own the model life-cycle end-to-end: data prep ➞ fine-tune/distil ➞ evaluate ➞ deploy ➞ monitor. Key Responsibilities Model Engineering • Fine-tune and quantise open-weight LLMs (Llama 3, Mistral, Gemma) and SLMs for low-latency edge inference. • Train or adapt computer-vision models (YOLO, Segment Anything, SAM-DINO) to detect site hazards, drawings anomalies, or asset states. Multimodal Pipelines • Build retrieval-augmented-generation (RAG) stacks: loaders → vector DB (FAISS / OpenSearch) → ranking prompts. • Combine vision + language outputs into single “scene → story” responses for dashboards and WhatsApp bots. Serving & MLOps • Package models as Docker images, SageMaker endpoints, or ONNX edge bundles; expose FastAPI/GRPC handlers with auth, rate-limit, telemetry. • Automate CI/CD: GitHub Actions → Terraform → blue-green deploys. Evaluation & Guardrails • Design automatic eval harnesses (BLEU, BERTScore, CLIP similarity, toxicity & bias checks). • Monitor drift, hallucination, latency; implement rollback triggers. Enablement & Storytelling • Write prompt playbooks & model cards so other teams can reuse your work. • Run internal workshops: “From design drawing to narrative” / “LLM safety by example”. Required Skills & Experience 3+ yrs ML/NLP/CV in production; at least 1 yr hands-on with Generative AI . Strong Python (FastAPI, Pydantic, asyncio) and HuggingFace Transformers OR diffusers . Experience with minimal-footprint models (LoRA, QLoRA, GGUF, INT-4) and vector search. Comfortable on AWS/GCP/Azure for GPU instances, serverless endpoints, IaC. Solid grasp of evaluation/guardrail frameworks (Helm, PromptLayer, Guardrails-AI, Triton metrics). Bonus Points Built a RAG or function-calling agent used by 500+ users. Prior CV pipeline (object-detection, segmentation) or speech-to-text real-time project. Live examples of creative prompt engineering or story-generation. Familiarity with LangChain, LlamaIndex, or BentoML. Why You’ll Love It Multidomain playground – text, vision, storytelling, decision-support. Tech freedom – pick the right model & stack; justify it; ship it. Remote-first – work anywhere ±4 hrs of IST; quarterly hack-weeks in Hyderabad. Top-quartile pay – base + milestone bonus + conference stipend. How to Apply Send a resume and link to GitHub / HF / Kaggle showcasing LLM or CV work. Include a 200-word note describing your favourite prompt or model tweak and the impact it had. Short-listed candidates complete a practical take-home (fine-tune tiny model, build RAG or vision demo, brief write-up) and a 45-min technical chat. We hire builders, not resume keywords. Show us you can ship AI that works in the real world—and explain it clearly—and you’re in. Show more Show less
Your mission in the first 12 months Fine-tune & play – take open-weight models (e.g. Mistral-7B-Instruct) and run LoRA or QLoRA experiments; report wins and trade-offs. Build data pipes – write small Python scripts that convert ERP tables, log streams, or image folders into tidy CSV / Parquet for training. Prototype RAG – load a batch of PDFs into a FAISS index, craft a prompt, and demo a chatbot that cites its sources. Model hygiene – add unit tests for prompt functions, set up basic latency / quality dashboards, hook OpenAI or Bedrock usage into CloudWatch. Team enablement – improve README files, jot quick snippets (“How to quantise with bits-and-bytes”), record Loom videos as you learn. What we’re looking for 0–2 yrs industry or strong capstone / internship in ML, NLP, or CV. Comfortable Python (Pandas, NumPy, basic PyTorch or TensorFlow). Some exposure to HuggingFace Transformers or OpenAI / Bedrock SDK. Knows how to use Git branches, open PRs, and write clear commit messages. Loves reading research blogs / papers and trying things fast. Bonus points if you have: a Kaggle medal, a Discord bot, or a small open-source repo that loads an LLM. Why you’ll like it here Real models in prod – your code will power WhatsApp workflows and dashboard narratives, not sit in a sandbox. Mentorship – pair weekly with a senior MCP Architect; monthly office hours with our Data-Science advisor. Tool budget – GPU credits, books, or a course of your choice each quarter. Hybrid freedom – work fully remote or drop into our Hyderabad office any time for whiteboard jams. How to apply Share a résumé and link to one small ML project (GitHub, Colab, Kaggle). Write 100 words on the coolest bug you solved while training or deploying a model. Short-listed candidates complete a two-hour take-home: fine-tune a tiny model or build a RAG proof-of-concept and explain your choices. Fresh minds + real models = fast progress. If you’re eager to learn and ship, let’s talk. Show more Show less
Job Opening – Junior AI-Agent Engineer (0–2 yrs • Full-Time • Remote-First) Why this role? We’re moving beyond single-shot prompts into autonomous agent chains that plan, reason and call tools (ERP APIs, SQL, CV models, WhatsApp bot) to finish real construction-industry tasks. If you’ve hacked together LangChain, CrewAI, or AutoGen projects and want to see them run in production, this is your playground. What you’ll tackle Prototype task-specific agents: • “DPR Validator” – reads today’s WhatsApp DPR, checks BOQ limits, writes feedback. • “Incident Triage Bot” – summarises image/video evidence, suggests severity and next steps. Wire agents to tools: Vector-DB RAG, ERPNext REST, SQL, Python calculators, WhatsApp send-API. Add memory & state (Redis / Postgres) so agents recall project history and user preferences. Write automated evals (truthfulness, task-success %, latency) and push results to Grafana. Package each agent as a FastAPI micro-service; deploy via GitHub Actions → EKS-Fargate. Document prompt chains, tool specs, and failure modes in Markdown + Mermaid. Must-have spark 0-2 yrs industry or a solid portfolio of LangChain / LlamaIndex / CrewAI / AutoGen hacks. Python 3 (async/await basics) and Git fluency (branch → PR → merge). Basic knowledge of OpenAI / Bedrock / HuggingFace APIs and one vector store (FAISS, Pinecone, OpenSearch). Comfortable reading REST docs and gluing APIs together. Clear written English; you can turn messy experiment notes into a crisp README. Bonus points Built a function-calling or tool-calling agent that actually shipped (Discord bot, Slack helper, side hustle). Exposure to Docker and simple CI (GitHub Actions). Familiar with evaluation frameworks (LangSmith, PromptLayer, ReAct eval notebooks). Success milestones Day 30: first agent demo hits WhatsApp sandbox, returns correct answer in < 5 s. Day 60: agent container auto-deploys with CI; dashboards show < 1 % failed calls. Day 90+: you add memory + feedback loop that improves task-success by ≥ 15 %. What we offer Remote-first, async-friendly culture; meet quarterly for hack-weeks (company-paid). Competitive junior salary + GPU credit + conference / course stipend. Direct mentorship from our MCP Architect and Applied-AI Lead—fast feedback loops. How to apply Send your resume and a link to at least one agent-style project (GitHub repo, Colab, or video demo). In ≤ 150 words, explain the coolest tool-calling or routing trick you built and what you learned.
About the Role You’ll join a small, fast team turning cutting-edge AI research into shippable products across text, vision, and multimodal domains. One sprint you’ll be distilling an LLM for WhatsApp chat-ops; the next you’ll be converting CAD drawings to BOM stories, or training a computer-vision model that flags onsite safety risks. You own the model life-cycle end-to-end: data prep ➞ fine-tune/distil ➞ evaluate ➞ deploy ➞ monitor. Key Responsibilities Model Engineering • Fine-tune and quantise open-weight LLMs (Llama 3, Mistral, Gemma) and SLMs for low-latency edge inference. • Train or adapt computer-vision models (YOLO, Segment Anything, SAM-DINO) to detect site hazards, drawings anomalies, or asset states. Multimodal Pipelines • Build retrieval-augmented-generation (RAG) stacks: loaders → vector DB (FAISS / OpenSearch) → ranking prompts. • Combine vision + language outputs into single “scene → story” responses for dashboards and WhatsApp bots. Serving & MLOps • Package models as Docker images, SageMaker endpoints, or ONNX edge bundles; expose FastAPI/GRPC handlers with auth, rate-limit, telemetry. • Automate CI/CD: GitHub Actions → Terraform → blue-green deploys. Evaluation & Guardrails • Design automatic eval harnesses (BLEU, BERTScore, CLIP similarity, toxicity & bias checks). • Monitor drift, hallucination, latency; implement rollback triggers. Enablement & Storytelling • Write prompt playbooks & model cards so other teams can reuse your work. • Run internal workshops: “From design drawing to narrative” / “LLM safety by example”. Required Skills & Experience 3+ yrs ML/NLP/CV in production; at least 1 yr hands-on with Generative AI . Strong Python (FastAPI, Pydantic, asyncio) and HuggingFace Transformers OR diffusers . Experience with minimal-footprint models (LoRA, QLoRA, GGUF, INT-4) and vector search. Comfortable on AWS/GCP/Azure for GPU instances, serverless endpoints, IaC. Solid grasp of evaluation/guardrail frameworks (Helm, PromptLayer, Guardrails-AI, Triton metrics). Bonus Points Built a RAG or function-calling agent used by 500+ users. Prior CV pipeline (object-detection, segmentation) or speech-to-text real-time project. Live examples of creative prompt engineering or story-generation. Familiarity with LangChain, LlamaIndex, or BentoML. Why You’ll Love It Multidomain playground – text, vision, storytelling, decision-support. Tech freedom – pick the right model & stack; justify it; ship it. Remote-first – work anywhere ±4 hrs of IST; quarterly hack-weeks in Hyderabad. Top-quartile pay – base + milestone bonus + conference stipend. How to Apply Send a resume and link to GitHub / HF / Kaggle showcasing LLM or CV work. Include a 200-word note describing your favourite prompt or model tweak and the impact it had. Short-listed candidates complete a practical take-home (fine-tune tiny model, build RAG or vision demo, brief write-up) and a 45-min technical chat. We hire builders, not resume keywords. Show us you can ship AI that works in the real world—and explain it clearly—and you’re in.
Your mission in the first 12 months Fine-tune & play – take open-weight models (e.g. Mistral-7B-Instruct) and run LoRA or QLoRA experiments; report wins and trade-offs. Build data pipes – write small Python scripts that convert ERP tables, log streams, or image folders into tidy CSV / Parquet for training. Prototype RAG – load a batch of PDFs into a FAISS index, craft a prompt, and demo a chatbot that cites its sources. Model hygiene – add unit tests for prompt functions, set up basic latency / quality dashboards, hook OpenAI or Bedrock usage into CloudWatch. Team enablement – improve README files, jot quick snippets (“How to quantise with bits-and-bytes”), record Loom videos as you learn. What we’re looking for 0–2 yrs industry or strong capstone / internship in ML, NLP, or CV. Comfortable Python (Pandas, NumPy, basic PyTorch or TensorFlow). Some exposure to HuggingFace Transformers or OpenAI / Bedrock SDK. Knows how to use Git branches, open PRs, and write clear commit messages. Loves reading research blogs / papers and trying things fast. Bonus points if you have: a Kaggle medal, a Discord bot, or a small open-source repo that loads an LLM. Why you’ll like it here Real models in prod – your code will power WhatsApp workflows and dashboard narratives, not sit in a sandbox. Mentorship – pair weekly with a senior MCP Architect; monthly office hours with our Data-Science advisor. Tool budget – GPU credits, books, or a course of your choice each quarter. Hybrid freedom – work fully remote or drop into our Hyderabad office any time for whiteboard jams. How to apply Share a résumé and link to one small ML project (GitHub, Colab, Kaggle). Write 100 words on the coolest bug you solved while training or deploying a model. Short-listed candidates complete a two-hour take-home: fine-tune a tiny model or build a RAG proof-of-concept and explain your choices. Fresh minds + real models = fast progress. If you’re eager to learn and ship, let’s talk.